Skip to content

O
opencv
Commit 7b87d72d authored by Vladislav Vinogradov's avatar Vladislav Vinogradov
Browse files
Options

refactored generalized hough (both CPU and GPU): 

removed set/get methods from Algorithm (implement owns)
removed GHT_* enumeration
parent 4b234fa0
Hide whitespace changes
Inline Side-by-side
Showing with 1028 additions and 1935 deletions
modules/gpuimgproc/doc/hough.rst
View file @ 7b87d72d
......@@ -213,98 +213,19 @@ Creates implementation for :ocv:class:`gpu::HoughCirclesDetector` .
gpu::GeneralizedHough
---------------------
.. ocv:class:: gpu::GeneralizedHough : public Algorithm
Base class for generalized hough transform. ::
class CV_EXPORTS GeneralizedHough : public Algorithm
{
public:
static Ptr<GeneralizedHough> create(int method);
virtual void setTemplate(InputArray templ, int cannyThreshold = 100, Point templCenter = Point(-1, -1)) = 0;
virtual void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter = Point(-1, -1)) = 0;
virtual void detect(InputArray image, OutputArray positions, int cannyThreshold = 100) = 0;
virtual void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions) = 0;
virtual void downloadResults(InputArray d_positions, OutputArray h_positions, OutputArray h_votes = noArray()) = 0;
};
Finds arbitrary template in the grayscale image using Generalized Hough Transform.
gpu::GeneralizedHough::create
-----------------------------
Creates implementation for :ocv:class:`gpu::GeneralizedHough` .
.. ocv:function:: Ptr<GeneralizedHough> gpu::GeneralizedHough::create(int method)
:param method: Combination of flags ( ``cv::GeneralizedHough::GHT_POSITION`` , ``cv::GeneralizedHough::GHT_SCALE`` , ``cv::GeneralizedHough::GHT_ROTATION`` ) specifying transformation to find.
For full affine transformations (move + scale + rotation) [Guil1999]_ algorithm is used, otherwise [Ballard1981]_ algorithm is used.
gpu::GeneralizedHough::setTemplate
gpu::createGeneralizedHoughBallard
----------------------------------
Set template to search.
.. ocv:function:: void gpu::GeneralizedHough::setTemplate(InputArray templ, int cannyThreshold = 100, Point templCenter = Point(-1, -1))
.. ocv:function:: void gpu::GeneralizedHough::setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter = Point(-1, -1))
Creates implementation for generalized hough transform from [Ballard1981]_ .
:param templ: Template image. Canny edge detector will be applied to extract template edges.
.. ocv:function:: Ptr<GeneralizedHoughBallard> gpu::createGeneralizedHoughBallard()
:param cannyThreshold: Threshold value for Canny edge detector.
:param templCenter: Center for rotation. By default image center will be used.
:param edges: Edge map for template image.
:param dx: First derivative of template image in the vertical direction. Support only ``CV_32S`` type.
:param dy: First derivative of template image in the horizontal direction. Support only ``CV_32S`` type.
gpu::GeneralizedHough::detect
-----------------------------
Finds template (set by :ocv:func:`gpu::GeneralizedHough::setTemplate` ) in the grayscale image.
.. ocv:function:: void gpu::GeneralizedHough::detect(InputArray image, OutputArray positions, int cannyThreshold = 100)
.. ocv:function:: void gpu::GeneralizedHough::detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions)
:param templ: Input image. Canny edge detector will be applied to extract template edges.
:param positions: Output vector of found objects. Each vector is encoded as a 4-element floating-point vector :math:`(x, y, scale, angle)` .
:param cannyThreshold: Threshold value for Canny edge detector.
:param edges: Edge map for input image.
:param dx: First derivative of input image in the vertical direction. Support only ``CV_32S`` type.
:param dy: First derivative of input image in the horizontal direction. Support only ``CV_32S`` type.
gpu::GeneralizedHough::downloadResults
--------------------------------------
Downloads results from :ocv:func:`gpu::GeneralizedHough::detect` to host memory.
.. ocv:function:: void gpu::GeneralizedHough::downloadResult(InputArray d_positions, OutputArray h_positions, OutputArray h_votes = noArray())
:param d_lines: Result of :ocv:func:`gpu::GeneralizedHough::detect` .
:param h_lines: Output host array.
gpu::createGeneralizedHoughGuil
-------------------------------
Creates implementation for generalized hough transform from [Guil1999]_ .
:param h_votes: Optional output array for votes. Each vector is encoded as a 3-element integer-point vector :math:`(position_votes, scale_votes, angle_votes)` .
.. ocv:function:: Ptr<GeneralizedHoughGuil> gpu::createGeneralizedHoughGuil()
......
modules/gpuimgproc/include/opencv2/gpuimgproc.hpp
View file @ 7b87d72d
......@@ -283,24 +283,13 @@ CV_EXPORTS Ptr<HoughCirclesDetector> createHoughCirclesDetector(float dp, float
//////////////////////////////////////
// GeneralizedHough
//! finds arbitrary template in the grayscale image using Generalized Hough Transform
//! Ballard, D.H. (1981). Generalizing the Hough transform to detect arbitrary shapes. Pattern Recognition 13 (2): 111-122.
//! Guil, N., González-Linares, J.M. and Zapata, E.L. (1999). Bidimensional shape detection using an invariant approach. Pattern Recognition 32 (6): 1025-1038.
class CV_EXPORTS GeneralizedHough : public Algorithm
{
public:
static Ptr<GeneralizedHough> create(int method);
//! set template to search
virtual void setTemplate(InputArray templ, int cannyThreshold = 100, Point templCenter = Point(-1, -1)) = 0;
virtual void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter = Point(-1, -1)) = 0;
//! Detects position only without traslation and rotation
CV_EXPORTS Ptr<GeneralizedHoughBallard> createGeneralizedHoughBallard();
//! find template on image
virtual void detect(InputArray image, OutputArray positions, int cannyThreshold = 100) = 0;
virtual void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions) = 0;
virtual void downloadResults(InputArray d_positions, OutputArray h_positions, OutputArray h_votes = noArray()) = 0;
};
//! Guil, N., González-Linares, J.M. and Zapata, E.L. (1999). Bidimensional shape detection using an invariant approach. Pattern Recognition 32 (6): 1025-1038.
//! Detects position, traslation and rotation
CV_EXPORTS Ptr<GeneralizedHoughGuil> createGeneralizedHoughGuil();
////////////////////////// Corners Detection ///////////////////////////
......
modules/gpuimgproc/perf/perf_hough.cpp
View file @ 7b87d72d
......@@ -227,23 +227,59 @@ PERF_TEST_P(Sz_Dp_MinDist, HoughCircles,
//////////////////////////////////////////////////////////////////////
// GeneralizedHough
enum { GHT_POSITION = cv::GeneralizedHough::GHT_POSITION,
GHT_SCALE = cv::GeneralizedHough::GHT_SCALE,
GHT_ROTATION = cv::GeneralizedHough::GHT_ROTATION
};
PERF_TEST_P(Sz, GeneralizedHoughBallard, GPU_TYPICAL_MAT_SIZES)
{
declare.time(10);
const cv::Size imageSize = GetParam();
const cv::Mat templ = readImage("cv/shared/templ.png", cv::IMREAD_GRAYSCALE);
ASSERT_FALSE(templ.empty());
cv::Mat image(imageSize, CV_8UC1, cv::Scalar::all(0));
templ.copyTo(image(cv::Rect(50, 50, templ.cols, templ.rows)));
cv::Mat edges;
cv::Canny(image, edges, 50, 100);
cv::Mat dx, dy;
cv::Sobel(image, dx, CV_32F, 1, 0);
cv::Sobel(image, dy, CV_32F, 0, 1);
if (PERF_RUN_GPU())
{
cv::Ptr<cv::GeneralizedHoughBallard> alg = cv::gpu::createGeneralizedHoughBallard();
const cv::gpu::GpuMat d_edges(edges);
const cv::gpu::GpuMat d_dx(dx);
const cv::gpu::GpuMat d_dy(dy);
cv::gpu::GpuMat positions;
alg->setTemplate(cv::gpu::GpuMat(templ));
TEST_CYCLE() alg->detect(d_edges, d_dx, d_dy, positions);
GPU_SANITY_CHECK(positions);
}
else
{
cv::Ptr<cv::GeneralizedHoughBallard> alg = cv::createGeneralizedHoughBallard();
cv::Mat positions;
alg->setTemplate(templ);
CV_FLAGS(GHMethod, GHT_POSITION, GHT_SCALE, GHT_ROTATION);
TEST_CYCLE() alg->detect(edges, dx, dy, positions);
DEF_PARAM_TEST(Method_Sz, GHMethod, cv::Size);
CPU_SANITY_CHECK(positions);
}
}
PERF_TEST_P(Method_Sz, GeneralizedHough,
Combine(Values(GHMethod(GHT_POSITION), GHMethod(GHT_POSITION | GHT_SCALE), GHMethod(GHT_POSITION | GHT_ROTATION), GHMethod(GHT_POSITION | GHT_SCALE | GHT_ROTATION)),
GPU_TYPICAL_MAT_SIZES))
PERF_TEST_P(Sz, GeneralizedHoughGuil, GPU_TYPICAL_MAT_SIZES)
{
declare.time(10);
const int method = GET_PARAM(0);
const cv::Size imageSize = GET_PARAM(1);
const cv::Size imageSize = GetParam();
const cv::Mat templ = readImage("cv/shared/templ.png", cv::IMREAD_GRAYSCALE);
ASSERT_FALSE(templ.empty());
......@@ -281,39 +317,32 @@ PERF_TEST_P(Method_Sz, GeneralizedHough,
if (PERF_RUN_GPU())
{
cv::Ptr<cv::GeneralizedHoughGuil> alg = cv::gpu::createGeneralizedHoughGuil();
alg->setMaxAngle(90.0);
alg->setAngleStep(2.0);
const cv::gpu::GpuMat d_edges(edges);
const cv::gpu::GpuMat d_dx(dx);
const cv::gpu::GpuMat d_dy(dy);
cv::gpu::GpuMat posAndVotes;
cv::gpu::GpuMat positions;
cv::Ptr<cv::gpu::GeneralizedHough> d_hough = cv::gpu::GeneralizedHough::create(method);
if (method & GHT_ROTATION)
{
d_hough->set("maxAngle", 90.0);
d_hough->set("angleStep", 2.0);
}
alg->setTemplate(cv::gpu::GpuMat(templ));
d_hough->setTemplate(cv::gpu::GpuMat(templ));
TEST_CYCLE() alg->detect(d_edges, d_dx, d_dy, positions);
TEST_CYCLE() d_hough->detect(d_edges, d_dx, d_dy, posAndVotes);
const cv::gpu::GpuMat positions(1, posAndVotes.cols, CV_32FC4, posAndVotes.data);
GPU_SANITY_CHECK(positions);
}
else
{
cv::Mat positions;
cv::Ptr<cv::GeneralizedHoughGuil> alg = cv::createGeneralizedHoughGuil();
alg->setMaxAngle(90.0);
alg->setAngleStep(2.0);
cv::Ptr<cv::GeneralizedHough> hough = cv::GeneralizedHough::create(method);
if (method & GHT_ROTATION)
{
hough->set("maxAngle", 90.0);
hough->set("angleStep", 2.0);
}
cv::Mat positions;
hough->setTemplate(templ);
alg->setTemplate(templ);
TEST_CYCLE() hough->detect(edges, dx, dy, positions);
TEST_CYCLE() alg->detect(edges, dx, dy, positions);
CPU_SANITY_CHECK(positions);
}
......
modules/gpuimgproc/src/cuda/generalized_hough.cu
View file @ 7b87d72d
......@@ -307,268 +307,6 @@ namespace cv { namespace gpu { namespace cudev
return totalCount;
}
////////////////////////////////////////////////////////////////////////
// Ballard_PosScale
__global__ void Ballard_PosScale_calcHist(const unsigned int* coordList, const float* thetaList,
PtrStep<short2> r_table, const int* r_sizes,
PtrStepi hist, const int rows, const int cols,
const float minScale, const float scaleStep, const int scaleRange,
const float idp, const float thetaScale)
{
const unsigned int coord = coordList[blockIdx.x];
float2 p;
p.x = (coord & 0xFFFF);
p.y = (coord >> 16) & 0xFFFF;
const float theta = thetaList[blockIdx.x];
const int n = __float2int_rn(theta * thetaScale);
const short2* r_row = r_table.ptr(n);
const int r_row_size = r_sizes[n];
for (int j = 0; j < r_row_size; ++j)
{
const float2 d = saturate_cast<float2>(r_row[j]);
for (int s = threadIdx.x; s < scaleRange; s += blockDim.x)
{
const float scale = minScale + s * scaleStep;
float2 c = p - scale * d;
c.x *= idp;
c.y *= idp;
if (c.x >= 0 && c.x < cols && c.y >= 0 && c.y < rows)
::atomicAdd(hist.ptr((s + 1) * (rows + 2) + __float2int_rn(c.y + 1)) + __float2int_rn(c.x + 1), 1);
}
}
}
void Ballard_PosScale_calcHist_gpu(const unsigned int* coordList, const float* thetaList, int pointsCount,
PtrStepSz<short2> r_table, const int* r_sizes,
PtrStepi hist, int rows, int cols,
float minScale, float scaleStep, int scaleRange,
float dp, int levels)
{
const dim3 block(256);
const dim3 grid(pointsCount);
const float idp = 1.0f / dp;
const float thetaScale = levels / (2.0f * CV_PI_F);
Ballard_PosScale_calcHist<<<grid, block>>>(coordList, thetaList,
r_table, r_sizes,
hist, rows, cols,
minScale, scaleStep, scaleRange,
idp, thetaScale);
cudaSafeCall( cudaGetLastError() );
cudaSafeCall( cudaDeviceSynchronize() );
}
__global__ void Ballard_PosScale_findPosInHist(const PtrStepi hist, const int rows, const int cols, const int scaleRange,
float4* out, int3* votes, const int maxSize,
const float minScale, const float scaleStep, const float dp, const int threshold)
{
const int x = blockIdx.x * blockDim.x + threadIdx.x;
const int y = blockIdx.y * blockDim.y + threadIdx.y;
if (x >= cols || y >= rows)
return;
for (int s = 0; s < scaleRange; ++s)
{
const float scale = minScale + s * scaleStep;
const int prevScaleIdx = (s) * (rows + 2);
const int curScaleIdx = (s + 1) * (rows + 2);
const int nextScaleIdx = (s + 2) * (rows + 2);
const int curVotes = hist(curScaleIdx + y + 1, x + 1);
if (curVotes > threshold &&
curVotes > hist(curScaleIdx + y + 1, x) &&
curVotes >= hist(curScaleIdx + y + 1, x + 2) &&
curVotes > hist(curScaleIdx + y, x + 1) &&
curVotes >= hist(curScaleIdx + y + 2, x + 1) &&
curVotes > hist(prevScaleIdx + y + 1, x + 1) &&
curVotes >= hist(nextScaleIdx + y + 1, x + 1))
{
const int ind = ::atomicAdd(&g_counter, 1);
if (ind < maxSize)
{
out[ind] = make_float4(x * dp, y * dp, scale, 0.0f);
votes[ind] = make_int3(curVotes, curVotes, 0);
}
}
}
}
int Ballard_PosScale_findPosInHist_gpu(PtrStepi hist, int rows, int cols, int scaleRange, float4* out, int3* votes, int maxSize,
float minScale, float scaleStep, float dp, int threshold)
{
void* counterPtr;
cudaSafeCall( cudaGetSymbolAddress(&counterPtr, g_counter) );
cudaSafeCall( cudaMemset(counterPtr, 0, sizeof(int)) );
const dim3 block(32, 8);
const dim3 grid(divUp(cols, block.x), divUp(rows, block.y));
cudaSafeCall( cudaFuncSetCacheConfig(Ballard_PosScale_findPosInHist, cudaFuncCachePreferL1) );
Ballard_PosScale_findPosInHist<<<grid, block>>>(hist, rows, cols, scaleRange, out, votes,
maxSize, minScale, scaleStep, dp, threshold);
cudaSafeCall( cudaGetLastError() );
cudaSafeCall( cudaDeviceSynchronize() );
int totalCount;
cudaSafeCall( cudaMemcpy(&totalCount, counterPtr, sizeof(int), cudaMemcpyDeviceToHost) );
totalCount = ::min(totalCount, maxSize);
return totalCount;
}
////////////////////////////////////////////////////////////////////////
// Ballard_PosRotation
__global__ void Ballard_PosRotation_calcHist(const unsigned int* coordList, const float* thetaList,
PtrStep<short2> r_table, const int* r_sizes,
PtrStepi hist, const int rows, const int cols,
const float minAngle, const float angleStep, const int angleRange,
const float idp, const float thetaScale)
{
const unsigned int coord = coordList[blockIdx.x];
float2 p;
p.x = (coord & 0xFFFF);
p.y = (coord >> 16) & 0xFFFF;
const float thetaVal = thetaList[blockIdx.x];
for (int a = threadIdx.x; a < angleRange; a += blockDim.x)
{
const float angle = (minAngle + a * angleStep) * (CV_PI_F / 180.0f);
float sinA, cosA;
sincosf(angle, &sinA, &cosA);
float theta = thetaVal - angle;
if (theta < 0)
theta += 2.0f * CV_PI_F;
const int n = __float2int_rn(theta * thetaScale);
const short2* r_row = r_table.ptr(n);
const int r_row_size = r_sizes[n];
for (int j = 0; j < r_row_size; ++j)
{
const float2 d = saturate_cast<float2>(r_row[j]);
const float2 dr = make_float2(d.x * cosA - d.y * sinA, d.x * sinA + d.y * cosA);
float2 c = make_float2(p.x - dr.x, p.y - dr.y);
c.x *= idp;
c.y *= idp;
if (c.x >= 0 && c.x < cols && c.y >= 0 && c.y < rows)
::atomicAdd(hist.ptr((a + 1) * (rows + 2) + __float2int_rn(c.y + 1)) + __float2int_rn(c.x + 1), 1);
}
}
}
void Ballard_PosRotation_calcHist_gpu(const unsigned int* coordList, const float* thetaList, int pointsCount,
PtrStepSz<short2> r_table, const int* r_sizes,
PtrStepi hist, int rows, int cols,
float minAngle, float angleStep, int angleRange,
float dp, int levels)
{
const dim3 block(256);
const dim3 grid(pointsCount);
const float idp = 1.0f / dp;
const float thetaScale = levels / (2.0f * CV_PI_F);
Ballard_PosRotation_calcHist<<<grid, block>>>(coordList, thetaList,
r_table, r_sizes,
hist, rows, cols,
minAngle, angleStep, angleRange,
idp, thetaScale);
cudaSafeCall( cudaGetLastError() );
cudaSafeCall( cudaDeviceSynchronize() );
}
__global__ void Ballard_PosRotation_findPosInHist(const PtrStepi hist, const int rows, const int cols, const int angleRange,
float4* out, int3* votes, const int maxSize,
const float minAngle, const float angleStep, const float dp, const int threshold)
{
const int x = blockIdx.x * blockDim.x + threadIdx.x;
const int y = blockIdx.y * blockDim.y + threadIdx.y;
if (x >= cols || y >= rows)
return;
for (int a = 0; a < angleRange; ++a)
{
const float angle = minAngle + a * angleStep;
const int prevAngleIdx = (a) * (rows + 2);
const int curAngleIdx = (a + 1) * (rows + 2);
const int nextAngleIdx = (a + 2) * (rows + 2);
const int curVotes = hist(curAngleIdx + y + 1, x + 1);
if (curVotes > threshold &&
curVotes > hist(curAngleIdx + y + 1, x) &&
curVotes >= hist(curAngleIdx + y + 1, x + 2) &&
curVotes > hist(curAngleIdx + y, x + 1) &&
curVotes >= hist(curAngleIdx + y + 2, x + 1) &&
curVotes > hist(prevAngleIdx + y + 1, x + 1) &&
curVotes >= hist(nextAngleIdx + y + 1, x + 1))
{
const int ind = ::atomicAdd(&g_counter, 1);
if (ind < maxSize)
{
out[ind] = make_float4(x * dp, y * dp, 1.0f, angle);
votes[ind] = make_int3(curVotes, 0, curVotes);
}
}
}
}
int Ballard_PosRotation_findPosInHist_gpu(PtrStepi hist, int rows, int cols, int angleRange, float4* out, int3* votes, int maxSize,
float minAngle, float angleStep, float dp, int threshold)
{
void* counterPtr;
cudaSafeCall( cudaGetSymbolAddress(&counterPtr, g_counter) );
cudaSafeCall( cudaMemset(counterPtr, 0, sizeof(int)) );
const dim3 block(32, 8);
const dim3 grid(divUp(cols, block.x), divUp(rows, block.y));
cudaSafeCall( cudaFuncSetCacheConfig(Ballard_PosRotation_findPosInHist, cudaFuncCachePreferL1) );
Ballard_PosRotation_findPosInHist<<<grid, block>>>(hist, rows, cols, angleRange, out, votes,
maxSize, minAngle, angleStep, dp, threshold);
cudaSafeCall( cudaGetLastError() );
cudaSafeCall( cudaDeviceSynchronize() );
int totalCount;
cudaSafeCall( cudaMemcpy(&totalCount, counterPtr, sizeof(int), cudaMemcpyDeviceToHost) );
totalCount = ::min(totalCount, maxSize);
return totalCount;
}
////////////////////////////////////////////////////////////////////////
// Guil_Full
......
modules/gpuimgproc/src/generalized_hough.cpp
View file @ 7b87d72d
......@@ -47,7 +47,9 @@ using namespace cv::gpu;
#if !defined (HAVE_CUDA) || defined (CUDA_DISABLER) || !defined(HAVE_OPENCV_GPUARITHM)
Ptr<gpu::GeneralizedHough> cv::gpu::GeneralizedHough::create(int) { throw_no_cuda(); return Ptr<GeneralizedHough>(); }
Ptr<GeneralizedHoughBallard> cv::gpu::createGeneralizedHoughBallard() { throw_no_cuda(); return Ptr<GeneralizedHoughBallard>(); }
Ptr<GeneralizedHoughGuil> cv::gpu::createGeneralizedHoughGuil() { throw_no_cuda(); return Ptr<GeneralizedHoughGuil>(); }
#else /* !defined (HAVE_CUDA) */
......@@ -67,22 +69,6 @@ namespace cv { namespace gpu { namespace cudev
float dp, int levels);
int Ballard_Pos_findPosInHist_gpu(PtrStepSzi hist, float4* out, int3* votes, int maxSize, float dp, int threshold);
void Ballard_PosScale_calcHist_gpu(const unsigned int* coordList, const float* thetaList, int pointsCount,
PtrStepSz<short2> r_table, const int* r_sizes,
PtrStepi hist, int rows, int cols,
float minScale, float scaleStep, int scaleRange,
float dp, int levels);
int Ballard_PosScale_findPosInHist_gpu(PtrStepi hist, int rows, int cols, int scaleRange, float4* out, int3* votes, int maxSize,
float minScale, float scaleStep, float dp, int threshold);
void Ballard_PosRotation_calcHist_gpu(const unsigned int* coordList, const float* thetaList, int pointsCount,
PtrStepSz<short2> r_table, const int* r_sizes,
PtrStepi hist, int rows, int cols,
float minAngle, float angleStep, int angleRange,
float dp, int levels);
int Ballard_PosRotation_findPosInHist_gpu(PtrStepi hist, int rows, int cols, int angleRange, float4* out, int3* votes, int maxSize,
float minAngle, float angleStep, float dp, int threshold);
void Guil_Full_setTemplFeatures(PtrStepb p1_pos, PtrStepb p1_theta, PtrStepb p2_pos, PtrStepb d12, PtrStepb r1, PtrStepb r2);
void Guil_Full_setImageFeatures(PtrStepb p1_pos, PtrStepb p1_theta, PtrStepb p2_pos, PtrStepb d12, PtrStepb r1, PtrStepb r2);
void Guil_Full_buildTemplFeatureList_gpu(const unsigned int* coordList, const float* thetaList, int pointsCount,
......@@ -110,278 +96,207 @@ namespace cv { namespace gpu { namespace cudev
}
}}}
// common
namespace
{
/////////////////////////////////////
// GeneralizedHoughBase
class GeneralizedHoughBase : public gpu::GeneralizedHough
class GeneralizedHoughBase
{
public:
protected:
GeneralizedHoughBase();
virtual ~GeneralizedHoughBase() {}
void setTemplate(InputArray templ, int cannyThreshold = 100, Point templCenter = Point(-1, -1));
void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter = Point(-1, -1));
void setTemplateImpl(InputArray templ, Point templCenter);
void setTemplateImpl(InputArray edges, InputArray dx, InputArray dy, Point templCenter);
void detect(InputArray image, OutputArray positions, int cannyThreshold = 100);
void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions);
void detectImpl(InputArray image, OutputArray positions, OutputArray votes);
void detectImpl(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes);
void downloadResults(InputArray d_positions, OutputArray h_positions, OutputArray h_votes = noArray());
void buildEdgePointList(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy);
protected:
virtual void setTemplateImpl(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy, Point templCenter) = 0;
virtual void detectImpl(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy, OutputArray positions) = 0;
virtual void processTempl() = 0;
virtual void processImage() = 0;
int cannyLowThresh_;
int cannyHighThresh_;
double minDist_;
double dp_;
int maxBufferSize_;
Size templSize_;
Point templCenter_;
GpuMat templEdges_;
GpuMat templDx_;
GpuMat templDy_;
Size imageSize_;
GpuMat imageEdges_;
GpuMat imageDx_;
GpuMat imageDy_;
GpuMat edgePointList_;
GpuMat outBuf_;
int posCount_;
private:
#ifdef HAVE_OPENCV_GPUFILTERS
GpuMat dx_, dy_;
GpuMat edges_;
void calcEdges(InputArray src, GpuMat& edges, GpuMat& dx, GpuMat& dy);
#endif
void filterMinDist();
void convertTo(OutputArray positions, OutputArray votes);
#ifdef HAVE_OPENCV_GPUFILTERS
Ptr<gpu::CannyEdgeDetector> canny_;
Ptr<gpu::Filter> filterDx_;
Ptr<gpu::Filter> filterDy_;
#endif
std::vector<float4> oldPosBuf_;
std::vector<int3> oldVoteBuf_;
std::vector<float4> newPosBuf_;
std::vector<int3> newVoteBuf_;
std::vector<int> indexies_;
};
GeneralizedHoughBase::GeneralizedHoughBase()
{
cannyLowThresh_ = 50;
cannyHighThresh_ = 100;
minDist_ = 1.0;
dp_ = 1.0;
maxBufferSize_ = 10000;
#ifdef HAVE_OPENCV_GPUFILTERS
canny_ = gpu::createCannyEdgeDetector(50, 100);
canny_ = gpu::createCannyEdgeDetector(cannyLowThresh_, cannyHighThresh_);
filterDx_ = gpu::createSobelFilter(CV_8UC1, CV_32S, 1, 0);
filterDy_ = gpu::createSobelFilter(CV_8UC1, CV_32S, 0, 1);
#endif
}
void GeneralizedHoughBase::setTemplate(InputArray _templ, int cannyThreshold, Point templCenter)
#ifdef HAVE_OPENCV_GPUFILTERS
void GeneralizedHoughBase::calcEdges(InputArray _src, GpuMat& edges, GpuMat& dx, GpuMat& dy)
{
#ifndef HAVE_OPENCV_GPUFILTERS
(void) _templ;
(void) cannyThreshold;
(void) templCenter;
throw_no_cuda();
#else
GpuMat templ = _templ.getGpuMat();
CV_Assert( templ.type() == CV_8UC1 );
CV_Assert( cannyThreshold > 0 );
GpuMat src = _src.getGpuMat();
ensureSizeIsEnough(templ.size(), CV_32SC1, dx_);
ensureSizeIsEnough(templ.size(), CV_32SC1, dy_);
CV_Assert( src.type() == CV_8UC1 );
CV_Assert( cannyLowThresh_ > 0 && cannyLowThresh_ < cannyHighThresh_ );
filterDx_->apply(templ, dx_);
filterDy_->apply(templ, dy_);
ensureSizeIsEnough(src.size(), CV_32SC1, dx);
ensureSizeIsEnough(src.size(), CV_32SC1, dy);
ensureSizeIsEnough(templ.size(), CV_8UC1, edges_);
filterDx_->apply(src, dx);
filterDy_->apply(src, dy);
canny_->setLowThreshold(cannyThreshold / 2);
canny_->setHighThreshold(cannyThreshold);
canny_->detect(dx_, dy_, edges_);
ensureSizeIsEnough(src.size(), CV_8UC1, edges);
if (templCenter == Point(-1, -1))
templCenter = Point(templ.cols / 2, templ.rows / 2);
setTemplateImpl(edges_, dx_, dy_, templCenter);
#endif
}
void GeneralizedHoughBase::setTemplate(InputArray _edges, InputArray _dx, InputArray _dy, Point templCenter)
{
GpuMat edges = _edges.getGpuMat();
GpuMat dx = _dx.getGpuMat();
GpuMat dy = _dy.getGpuMat();
if (templCenter == Point(-1, -1))
templCenter = Point(edges.cols / 2, edges.rows / 2);
setTemplateImpl(edges, dx, dy, templCenter);
canny_->setLowThreshold(cannyLowThresh_);
canny_->setHighThreshold(cannyHighThresh_);
canny_->detect(dx, dy, edges);
}
#endif
void GeneralizedHoughBase::detect(InputArray _image, OutputArray positions, int cannyThreshold)
void GeneralizedHoughBase::setTemplateImpl(InputArray templ, Point templCenter)
{
#ifndef HAVE_OPENCV_GPUFILTERS
(void) _image;
(void) positions;
(void) cannyThreshold;
(void) templ;
(void) templCenter;
throw_no_cuda();
#else
GpuMat image = _image.getGpuMat();
CV_Assert( image.type() == CV_8UC1 );
CV_Assert( cannyThreshold > 0 );
ensureSizeIsEnough(image.size(), CV_32SC1, dx_);
ensureSizeIsEnough(image.size(), CV_32SC1, dy_);
filterDx_->apply(image, dx_);
filterDy_->apply(image, dy_);
calcEdges(templ, templEdges_, templDx_, templDy_);
ensureSizeIsEnough(image.size(), CV_8UC1, edges_);
if (templCenter == Point(-1, -1))
templCenter = Point(templEdges_.cols / 2, templEdges_.rows / 2);
canny_->setLowThreshold(cannyThreshold / 2);
canny_->setHighThreshold(cannyThreshold);
canny_->detect(dx_, dy_, edges_);
templSize_ = templEdges_.size();
templCenter_ = templCenter;
detectImpl(edges_, dx_, dy_, positions);
processTempl();
#endif
}
void GeneralizedHoughBase::detect(InputArray _edges, InputArray _dx, InputArray _dy, OutputArray positions)
void GeneralizedHoughBase::setTemplateImpl(InputArray edges, InputArray dx, InputArray dy, Point templCenter)
{
GpuMat edges = _edges.getGpuMat();
GpuMat dx = _dx.getGpuMat();
GpuMat dy = _dy.getGpuMat();
detectImpl(edges, dx, dy, positions);
}
void GeneralizedHoughBase::downloadResults(InputArray _d_positions, OutputArray h_positions, OutputArray h_votes)
{
GpuMat d_positions = _d_positions.getGpuMat();
if (d_positions.empty())
{
h_positions.release();
if (h_votes.needed())
h_votes.release();
return;
}
edges.getGpuMat().copyTo(templEdges_);
dx.getGpuMat().copyTo(templDx_);
dy.getGpuMat().copyTo(templDy_);
CV_Assert( d_positions.rows == 2 && d_positions.type() == CV_32FC4 );
CV_Assert( templEdges_.type() == CV_8UC1 );
CV_Assert( templDx_.type() == CV_32FC1 && templDx_.size() == templEdges_.size() );
CV_Assert( templDy_.type() == templDx_.type() && templDy_.size() == templEdges_.size() );
d_positions.row(0).download(h_positions);
if (h_votes.needed())
{
GpuMat d_votes(1, d_positions.cols, CV_32SC3, d_positions.ptr<int3>(1));
d_votes.download(h_votes);
}
}
if (templCenter == Point(-1, -1))
templCenter = Point(templEdges_.cols / 2, templEdges_.rows / 2);
/////////////////////////////////////
// GHT_Pos
templSize_ = templEdges_.size();
templCenter_ = templCenter;
template <typename T, class A> void releaseVector(std::vector<T, A>& v)
{
std::vector<T, A> empty;
empty.swap(v);
processTempl();
}
class GHT_Pos : public GeneralizedHoughBase
void GeneralizedHoughBase::detectImpl(InputArray image, OutputArray positions, OutputArray votes)
{
public:
GHT_Pos();
protected:
void setTemplateImpl(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy, Point templCenter);
void detectImpl(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy, OutputArray positions);
void releaseImpl();
virtual void processTempl() = 0;
virtual void processImage() = 0;
void buildEdgePointList(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy);
void filterMinDist();
void convertTo(OutputArray positions);
int maxSize;
double minDist;
Size templSize;
Point templCenter;
GpuMat templEdges;
GpuMat templDx;
GpuMat templDy;
Size imageSize;
GpuMat imageEdges;
GpuMat imageDx;
GpuMat imageDy;
GpuMat edgePointList;
GpuMat outBuf;
int posCount;
std::vector<float4> oldPosBuf;
std::vector<int3> oldVoteBuf;
std::vector<float4> newPosBuf;
std::vector<int3> newVoteBuf;
std::vector<int> indexies;
};
GHT_Pos::GHT_Pos()
{
maxSize = 10000;
minDist = 1.0;
}
#ifndef HAVE_OPENCV_GPUFILTERS
(void) templ;
(void) templCenter;
throw_no_cuda();
#else
calcEdges(image, imageEdges_, imageDx_, imageDy_);
void GHT_Pos::setTemplateImpl(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy, Point templCenter_)
{
templSize = edges.size();
templCenter = templCenter_;
imageSize_ = imageEdges_.size();
ensureSizeIsEnough(templSize, edges.type(), templEdges);
ensureSizeIsEnough(templSize, dx.type(), templDx);
ensureSizeIsEnough(templSize, dy.type(), templDy);
posCount_ = 0;
edges.copyTo(templEdges);
dx.copyTo(templDx);
dy.copyTo(templDy);
processImage();
processTempl();
if (posCount_ == 0)
{
positions.release();
if (votes.needed())
votes.release();
}
else
{
if (minDist_ > 1)
filterMinDist();
convertTo(positions, votes);
}
#endif
}
void GHT_Pos::detectImpl(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy, OutputArray positions)
void GeneralizedHoughBase::detectImpl(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes)
{
imageSize = edges.size();
edges.getGpuMat().copyTo(imageEdges_);
dx.getGpuMat().copyTo(imageDx_);
dy.getGpuMat().copyTo(imageDy_);
ensureSizeIsEnough(imageSize, edges.type(), imageEdges);
ensureSizeIsEnough(imageSize, dx.type(), imageDx);
ensureSizeIsEnough(imageSize, dy.type(), imageDy);
CV_Assert( imageEdges_.type() == CV_8UC1 );
CV_Assert( imageDx_.type() == CV_32FC1 && imageDx_.size() == imageEdges_.size() );
CV_Assert( imageDy_.type() == imageDx_.type() && imageDy_.size() == imageEdges_.size() );
edges.copyTo(imageEdges);
dx.copyTo(imageDx);
dy.copyTo(imageDy);
imageSize_ = imageEdges_.size();
posCount = 0;
posCount_ = 0;
processImage();
if (posCount == 0)
if (posCount_ == 0)
{
positions.release();
if (votes.needed())
votes.release();
}
else
{
if (minDist > 1)
if (minDist_ > 1)
filterMinDist();
convertTo(positions);
convertTo(positions, votes);
}
}
void GHT_Pos::releaseImpl()
{
templSize = Size();
templCenter = Point(-1, -1);
templEdges.release();
templDx.release();
templDy.release();
imageSize = Size();
imageEdges.release();
imageDx.release();
imageDy.release();
edgePointList.release();
outBuf.release();
posCount = 0;
releaseVector(oldPosBuf);
releaseVector(oldVoteBuf);
releaseVector(newPosBuf);
releaseVector(newVoteBuf);
releaseVector(indexies);
}
void GHT_Pos::buildEdgePointList(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy)
void GeneralizedHoughBase::buildEdgePointList(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy)
{
using namespace cv::gpu::cudev::ght;
......@@ -397,17 +312,17 @@ namespace
0
};
CV_Assert(edges.type() == CV_8UC1);
CV_Assert(dx.size() == edges.size());
CV_Assert(dy.type() == dx.type() && dy.size() == edges.size());
CV_Assert( edges.type() == CV_8UC1 );
CV_Assert( dx.size() == edges.size() );
CV_Assert( dy.type() == dx.type() && dy.size() == edges.size() );
const func_t func = funcs[dx.depth()];
CV_Assert(func != 0);
CV_Assert( func != 0 );
edgePointList.cols = (int) (edgePointList.step / sizeof(int));
ensureSizeIsEnough(2, edges.size().area(), CV_32SC1, edgePointList);
edgePointList_.cols = (int) (edgePointList_.step / sizeof(int));
ensureSizeIsEnough(2, edges.size().area(), CV_32SC1, edgePointList_);
edgePointList.cols = func(edges, dx, dy, edgePointList.ptr<unsigned int>(0), edgePointList.ptr<float>(1));
edgePointList_.cols = func(edges, dx, dy, edgePointList_.ptr<unsigned int>(0), edgePointList_.ptr<float>(1));
}
struct IndexCmp
......@@ -422,37 +337,37 @@ namespace
}
};
void GHT_Pos::filterMinDist()
void GeneralizedHoughBase::filterMinDist()
{
oldPosBuf.resize(posCount);
oldVoteBuf.resize(posCount);
oldPosBuf_.resize(posCount_);
oldVoteBuf_.resize(posCount_);
cudaSafeCall( cudaMemcpy(&oldPosBuf[0], outBuf.ptr(0), posCount * sizeof(float4), cudaMemcpyDeviceToHost) );
cudaSafeCall( cudaMemcpy(&oldVoteBuf[0], outBuf.ptr(1), posCount * sizeof(int3), cudaMemcpyDeviceToHost) );
cudaSafeCall( cudaMemcpy(&oldPosBuf_[0], outBuf_.ptr(0), posCount_ * sizeof(float4), cudaMemcpyDeviceToHost) );
cudaSafeCall( cudaMemcpy(&oldVoteBuf_[0], outBuf_.ptr(1), posCount_ * sizeof(int3), cudaMemcpyDeviceToHost) );
indexies.resize(posCount);
for (int i = 0; i < posCount; ++i)
indexies[i] = i;
std::sort(indexies.begin(), indexies.end(), IndexCmp(&oldVoteBuf[0]));
indexies_.resize(posCount_);
for (int i = 0; i < posCount_; ++i)
indexies_[i] = i;
std::sort(indexies_.begin(), indexies_.end(), IndexCmp(&oldVoteBuf_[0]));
newPosBuf.clear();
newVoteBuf.clear();
newPosBuf.reserve(posCount);
newVoteBuf.reserve(posCount);
newPosBuf_.clear();
newVoteBuf_.clear();
newPosBuf_.reserve(posCount_);
newVoteBuf_.reserve(posCount_);
const int cellSize = cvRound(minDist);
const int gridWidth = (imageSize.width + cellSize - 1) / cellSize;
const int gridHeight = (imageSize.height + cellSize - 1) / cellSize;
const int cellSize = cvRound(minDist_);
const int gridWidth = (imageSize_.width + cellSize - 1) / cellSize;
const int gridHeight = (imageSize_.height + cellSize - 1) / cellSize;
std::vector< std::vector<Point2f> > grid(gridWidth * gridHeight);
const double minDist2 = minDist * minDist;
const double minDist2 = minDist_ * minDist_;
for (int i = 0; i < posCount; ++i)
for (int i = 0; i < posCount_; ++i)
{
const int ind = indexies[i];
const int ind = indexies_[i];
Point2f p(oldPosBuf[ind].x, oldPosBuf[ind].y);
Point2f p(oldPosBuf_[ind].x, oldPosBuf_[ind].y);
bool good = true;
......@@ -495,353 +410,238 @@ namespace
{
grid[yCell * gridWidth + xCell].push_back(p);
newPosBuf.push_back(oldPosBuf[ind]);
newVoteBuf.push_back(oldVoteBuf[ind]);
newPosBuf_.push_back(oldPosBuf_[ind]);
newVoteBuf_.push_back(oldVoteBuf_[ind]);
}
}
posCount = static_cast<int>(newPosBuf.size());
cudaSafeCall( cudaMemcpy(outBuf.ptr(0), &newPosBuf[0], posCount * sizeof(float4), cudaMemcpyHostToDevice) );
cudaSafeCall( cudaMemcpy(outBuf.ptr(1), &newVoteBuf[0], posCount * sizeof(int3), cudaMemcpyHostToDevice) );
posCount_ = static_cast<int>(newPosBuf_.size());
cudaSafeCall( cudaMemcpy(outBuf_.ptr(0), &newPosBuf_[0], posCount_ * sizeof(float4), cudaMemcpyHostToDevice) );
cudaSafeCall( cudaMemcpy(outBuf_.ptr(1), &newVoteBuf_[0], posCount_ * sizeof(int3), cudaMemcpyHostToDevice) );
}
void GHT_Pos::convertTo(OutputArray positions)
void GeneralizedHoughBase::convertTo(OutputArray positions, OutputArray votes)
{
ensureSizeIsEnough(2, posCount, CV_32FC4, positions);
GpuMat(2, posCount, CV_32FC4, outBuf.data, outBuf.step).copyTo(positions);
ensureSizeIsEnough(1, posCount_, CV_32FC4, positions);
GpuMat(1, posCount_, CV_32FC4, outBuf_.ptr(0), outBuf_.step).copyTo(positions);
if (votes.needed())
{
ensureSizeIsEnough(1, posCount_, CV_32FC3, votes);
GpuMat(1, posCount_, CV_32FC4, outBuf_.ptr(1), outBuf_.step).copyTo(votes);
}
}
}
/////////////////////////////////////
// POSITION Ballard
// GeneralizedHoughBallard
class GHT_Ballard_Pos : public GHT_Pos
namespace
{
class GeneralizedHoughBallardImpl : public GeneralizedHoughBallard, private GeneralizedHoughBase
{
public:
AlgorithmInfo* info() const;
GHT_Ballard_Pos();
protected:
void releaseImpl();
GeneralizedHoughBallardImpl();
void processTempl();
void processImage();
void setTemplate(InputArray templ, Point templCenter) { setTemplateImpl(templ, templCenter); }
void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter) { setTemplateImpl(edges, dx, dy, templCenter); }
virtual void calcHist();
virtual void findPosInHist();
void detect(InputArray image, OutputArray positions, OutputArray votes) { detectImpl(image, positions, votes); }
void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes) { detectImpl(edges, dx, dy, positions, votes); }
int levels;
int votesThreshold;
double dp;
void setCannyLowThresh(int cannyLowThresh) { cannyLowThresh_ = cannyLowThresh; }
int getCannyLowThresh() const { return cannyLowThresh_; }
GpuMat r_table;
GpuMat r_sizes;
void setCannyHighThresh(int cannyHighThresh) { cannyHighThresh_ = cannyHighThresh; }
int getCannyHighThresh() const { return cannyHighThresh_; }
GpuMat hist;
};
void setMinDist(double minDist) { minDist_ = minDist; }
double getMinDist() const { return minDist_; }
CV_INIT_ALGORITHM(GHT_Ballard_Pos, "GeneralizedHough_GPU.POSITION",
obj.info()->addParam(obj, "maxSize", obj.maxSize, false, 0, 0,
"Maximal size of inner buffers.");
obj.info()->addParam(obj, "minDist", obj.minDist, false, 0, 0,
"Minimum distance between the centers of the detected objects.");
obj.info()->addParam(obj, "levels", obj.levels, false, 0, 0,
"R-Table levels.");
obj.info()->addParam(obj, "votesThreshold", obj.votesThreshold, false, 0, 0,
"The accumulator threshold for the template centers at the detection stage. The smaller it is, the more false positions may be detected.");
obj.info()->addParam(obj, "dp", obj.dp, false, 0, 0,
"Inverse ratio of the accumulator resolution to the image resolution."));
GHT_Ballard_Pos::GHT_Ballard_Pos()
{
levels = 360;
votesThreshold = 100;
dp = 1.0;
}
void setDp(double dp) { dp_ = dp; }
double getDp() const { return dp_; }
void GHT_Ballard_Pos::releaseImpl()
{
GHT_Pos::releaseImpl();
void setMaxBufferSize(int maxBufferSize) { maxBufferSize_ = maxBufferSize; }
int getMaxBufferSize() const { return maxBufferSize_; }
r_table.release();
r_sizes.release();
void setLevels(int levels) { levels_ = levels; }
int getLevels() const { return levels_; }
hist.release();
}
void setVotesThreshold(int votesThreshold) { votesThreshold_ = votesThreshold; }
int getVotesThreshold() const { return votesThreshold_; }
void GHT_Ballard_Pos::processTempl()
{
using namespace cv::gpu::cudev::ght;
private:
void processTempl();
void processImage();
CV_Assert(levels > 0);
void calcHist();
void findPosInHist();
buildEdgePointList(templEdges, templDx, templDy);
int levels_;
int votesThreshold_;
ensureSizeIsEnough(levels + 1, maxSize, CV_16SC2, r_table);
ensureSizeIsEnough(1, levels + 1, CV_32SC1, r_sizes);
r_sizes.setTo(Scalar::all(0));
GpuMat r_table_;
GpuMat r_sizes_;
if (edgePointList.cols > 0)
{
buildRTable_gpu(edgePointList.ptr<unsigned int>(0), edgePointList.ptr<float>(1), edgePointList.cols,
r_table, r_sizes.ptr<int>(), make_short2(templCenter.x, templCenter.y), levels);
gpu::min(r_sizes, maxSize, r_sizes);
}
}
GpuMat hist_;
};
void GHT_Ballard_Pos::processImage()
GeneralizedHoughBallardImpl::GeneralizedHoughBallardImpl()
{
calcHist();
findPosInHist();
levels_ = 360;
votesThreshold_ = 100;
}
void GHT_Ballard_Pos::calcHist()
void GeneralizedHoughBallardImpl::processTempl()
{
using namespace cv::gpu::cudev::ght;
CV_Assert(levels > 0 && r_table.rows == (levels + 1) && r_sizes.cols == (levels + 1));
CV_Assert(dp > 0.0);
const double idp = 1.0 / dp;
CV_Assert( levels_ > 0 );
buildEdgePointList(imageEdges, imageDx, imageDy);
buildEdgePointList(templEdges_, templDx_, templDy_);
ensureSizeIsEnough(cvCeil(imageSize.height * idp) + 2, cvCeil(imageSize.width * idp) + 2, CV_32SC1, hist);
hist.setTo(Scalar::all(0));
ensureSizeIsEnough(levels_ + 1, maxBufferSize_, CV_16SC2, r_table_);
ensureSizeIsEnough(1, levels_ + 1, CV_32SC1, r_sizes_);
r_sizes_.setTo(Scalar::all(0));
if (edgePointList.cols > 0)
if (edgePointList_.cols > 0)
{
Ballard_Pos_calcHist_gpu(edgePointList.ptr<unsigned int>(0), edgePointList.ptr<float>(1), edgePointList.cols,
r_table, r_sizes.ptr<int>(),
hist,
(float)dp, levels);
buildRTable_gpu(edgePointList_.ptr<unsigned int>(0), edgePointList_.ptr<float>(1), edgePointList_.cols,
r_table_, r_sizes_.ptr<int>(), make_short2(templCenter_.x, templCenter_.y), levels_);
gpu::min(r_sizes_, maxBufferSize_, r_sizes_);
}
}
void GHT_Ballard_Pos::findPosInHist()
void GeneralizedHoughBallardImpl::processImage()
{
using namespace cv::gpu::cudev::ght;
CV_Assert(votesThreshold > 0);
ensureSizeIsEnough(2, maxSize, CV_32FC4, outBuf);
posCount = Ballard_Pos_findPosInHist_gpu(hist, outBuf.ptr<float4>(0), outBuf.ptr<int3>(1), maxSize, (float)dp, votesThreshold);
}
/////////////////////////////////////
// POSITION & SCALE
class GHT_Ballard_PosScale : public GHT_Ballard_Pos
{
public:
AlgorithmInfo* info() const;
GHT_Ballard_PosScale();
protected:
void calcHist();
void findPosInHist();
double minScale;
double maxScale;
double scaleStep;
};
CV_INIT_ALGORITHM(GHT_Ballard_PosScale, "GeneralizedHough_GPU.POSITION_SCALE",
obj.info()->addParam(obj, "maxSize", obj.maxSize, false, 0, 0,
"Maximal size of inner buffers.");
obj.info()->addParam(obj, "minDist", obj.minDist, false, 0, 0,
"Minimum distance between the centers of the detected objects.");
obj.info()->addParam(obj, "levels", obj.levels, false, 0, 0,
"R-Table levels.");
obj.info()->addParam(obj, "votesThreshold", obj.votesThreshold, false, 0, 0,
"The accumulator threshold for the template centers at the detection stage. The smaller it is, the more false positions may be detected.");
obj.info()->addParam(obj, "dp", obj.dp, false, 0, 0,
"Inverse ratio of the accumulator resolution to the image resolution.");
obj.info()->addParam(obj, "minScale", obj.minScale, false, 0, 0,
"Minimal scale to detect.");
obj.info()->addParam(obj, "maxScale", obj.maxScale, false, 0, 0,
"Maximal scale to detect.");
obj.info()->addParam(obj, "scaleStep", obj.scaleStep, false, 0, 0,
"Scale step."));
GHT_Ballard_PosScale::GHT_Ballard_PosScale()
{
minScale = 0.5;
maxScale = 2.0;
scaleStep = 0.05;
calcHist();
findPosInHist();
}
void GHT_Ballard_PosScale::calcHist()
void GeneralizedHoughBallardImpl::calcHist()
{
using namespace cv::gpu::cudev::ght;
CV_Assert(levels > 0 && r_table.rows == (levels + 1) && r_sizes.cols == (levels + 1));
CV_Assert(dp > 0.0);
CV_Assert(minScale > 0.0 && minScale < maxScale);
CV_Assert(scaleStep > 0.0);
CV_Assert( levels_ > 0 && r_table_.rows == (levels_ + 1) && r_sizes_.cols == (levels_ + 1) );
CV_Assert( dp_ > 0.0);
const double idp = 1.0 / dp;
const int scaleRange = cvCeil((maxScale - minScale) / scaleStep);
const int rows = cvCeil(imageSize.height * idp);
const int cols = cvCeil(imageSize.width * idp);
const double idp = 1.0 / dp_;
buildEdgePointList(imageEdges, imageDx, imageDy);
buildEdgePointList(imageEdges_, imageDx_, imageDy_);
ensureSizeIsEnough((scaleRange + 2) * (rows + 2), cols + 2, CV_32SC1, hist);
hist.setTo(Scalar::all(0));
ensureSizeIsEnough(cvCeil(imageSize_.height * idp) + 2, cvCeil(imageSize_.width * idp) + 2, CV_32SC1, hist_);
hist_.setTo(Scalar::all(0));
if (edgePointList.cols > 0)
if (edgePointList_.cols > 0)
{
Ballard_PosScale_calcHist_gpu(edgePointList.ptr<unsigned int>(0), edgePointList.ptr<float>(1), edgePointList.cols,
r_table, r_sizes.ptr<int>(),
hist, rows, cols,
(float)minScale, (float)scaleStep, scaleRange, (float)dp, levels);
Ballard_Pos_calcHist_gpu(edgePointList_.ptr<unsigned int>(0), edgePointList_.ptr<float>(1), edgePointList_.cols,
r_table_, r_sizes_.ptr<int>(),
hist_,
(float)dp_, levels_);
}
}
void GHT_Ballard_PosScale::findPosInHist()
void GeneralizedHoughBallardImpl::findPosInHist()
{
using namespace cv::gpu::cudev::ght;
CV_Assert(votesThreshold > 0);
const double idp = 1.0 / dp;
const int scaleRange = cvCeil((maxScale - minScale) / scaleStep);
const int rows = cvCeil(imageSize.height * idp);
const int cols = cvCeil(imageSize.width * idp);
CV_Assert( votesThreshold_ > 0 );
ensureSizeIsEnough(2, maxSize, CV_32FC4, outBuf);
ensureSizeIsEnough(2, maxBufferSize_, CV_32FC4, outBuf_);
posCount = Ballard_PosScale_findPosInHist_gpu(hist, rows, cols, scaleRange, outBuf.ptr<float4>(0), outBuf.ptr<int3>(1), maxSize, (float)minScale, (float)scaleStep, (float)dp, votesThreshold);
posCount_ = Ballard_Pos_findPosInHist_gpu(hist_, outBuf_.ptr<float4>(0), outBuf_.ptr<int3>(1), maxBufferSize_, (float)dp_, votesThreshold_);
}
}
/////////////////////////////////////
// POSITION & Rotation
Ptr<GeneralizedHoughBallard> cv::gpu::createGeneralizedHoughBallard()
{
return new GeneralizedHoughBallardImpl;
}
class GHT_Ballard_PosRotation : public GHT_Ballard_Pos
// GeneralizedHoughGuil
namespace
{
class GeneralizedHoughGuilImpl : public GeneralizedHoughGuil, private GeneralizedHoughBase
{
public:
AlgorithmInfo* info() const;
GeneralizedHoughGuilImpl();
GHT_Ballard_PosRotation();
void setTemplate(InputArray templ, Point templCenter) { setTemplateImpl(templ, templCenter); }
void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter) { setTemplateImpl(edges, dx, dy, templCenter); }
protected:
void calcHist();
void findPosInHist();
void detect(InputArray image, OutputArray positions, OutputArray votes) { detectImpl(image, positions, votes); }
void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes) { detectImpl(edges, dx, dy, positions, votes); }
double minAngle;
double maxAngle;
double angleStep;
};
void setCannyLowThresh(int cannyLowThresh) { cannyLowThresh_ = cannyLowThresh; }
int getCannyLowThresh() const { return cannyLowThresh_; }
CV_INIT_ALGORITHM(GHT_Ballard_PosRotation, "GeneralizedHough_GPU.POSITION_ROTATION",
obj.info()->addParam(obj, "maxSize", obj.maxSize, false, 0, 0,
"Maximal size of inner buffers.");
obj.info()->addParam(obj, "minDist", obj.minDist, false, 0, 0,
"Minimum distance between the centers of the detected objects.");
obj.info()->addParam(obj, "levels", obj.levels, false, 0, 0,
"R-Table levels.");
obj.info()->addParam(obj, "votesThreshold", obj.votesThreshold, false, 0, 0,
"The accumulator threshold for the template centers at the detection stage. The smaller it is, the more false positions may be detected.");
obj.info()->addParam(obj, "dp", obj.dp, false, 0, 0,
"Inverse ratio of the accumulator resolution to the image resolution.");
obj.info()->addParam(obj, "minAngle", obj.minAngle, false, 0, 0,
"Minimal rotation angle to detect in degrees.");
obj.info()->addParam(obj, "maxAngle", obj.maxAngle, false, 0, 0,
"Maximal rotation angle to detect in degrees.");
obj.info()->addParam(obj, "angleStep", obj.angleStep, false, 0, 0,
"Angle step in degrees."));
GHT_Ballard_PosRotation::GHT_Ballard_PosRotation()
{
minAngle = 0.0;
maxAngle = 360.0;
angleStep = 1.0;
}
void setCannyHighThresh(int cannyHighThresh) { cannyHighThresh_ = cannyHighThresh; }
int getCannyHighThresh() const { return cannyHighThresh_; }
void GHT_Ballard_PosRotation::calcHist()
{
using namespace cv::gpu::cudev::ght;
void setMinDist(double minDist) { minDist_ = minDist; }
double getMinDist() const { return minDist_; }
CV_Assert(levels > 0 && r_table.rows == (levels + 1) && r_sizes.cols == (levels + 1));
CV_Assert(dp > 0.0);
CV_Assert(minAngle >= 0.0 && minAngle < maxAngle && maxAngle <= 360.0);
CV_Assert(angleStep > 0.0 && angleStep < 360.0);
void setDp(double dp) { dp_ = dp; }
double getDp() const { return dp_; }
const double idp = 1.0 / dp;
const int angleRange = cvCeil((maxAngle - minAngle) / angleStep);
const int rows = cvCeil(imageSize.height * idp);
const int cols = cvCeil(imageSize.width * idp);
void setMaxBufferSize(int maxBufferSize) { maxBufferSize_ = maxBufferSize; }
int getMaxBufferSize() const { return maxBufferSize_; }
buildEdgePointList(imageEdges, imageDx, imageDy);
void setXi(double xi) { xi_ = xi; }
double getXi() const { return xi_; }
ensureSizeIsEnough((angleRange + 2) * (rows + 2), cols + 2, CV_32SC1, hist);
hist.setTo(Scalar::all(0));
void setLevels(int levels) { levels_ = levels; }
int getLevels() const { return levels_; }
if (edgePointList.cols > 0)
{
Ballard_PosRotation_calcHist_gpu(edgePointList.ptr<unsigned int>(0), edgePointList.ptr<float>(1), edgePointList.cols,
r_table, r_sizes.ptr<int>(),
hist, rows, cols,
(float)minAngle, (float)angleStep, angleRange, (float)dp, levels);
}
}
void setAngleEpsilon(double angleEpsilon) { angleEpsilon_ = angleEpsilon; }
double getAngleEpsilon() const { return angleEpsilon_; }
void GHT_Ballard_PosRotation::findPosInHist()
{
using namespace cv::gpu::cudev::ght;
void setMinAngle(double minAngle) { minAngle_ = minAngle; }
double getMinAngle() const { return minAngle_; }
CV_Assert(votesThreshold > 0);
void setMaxAngle(double maxAngle) { maxAngle_ = maxAngle; }
double getMaxAngle() const { return maxAngle_; }
const double idp = 1.0 / dp;
const int angleRange = cvCeil((maxAngle - minAngle) / angleStep);
const int rows = cvCeil(imageSize.height * idp);
const int cols = cvCeil(imageSize.width * idp);
void setAngleStep(double angleStep) { angleStep_ = angleStep; }
double getAngleStep() const { return angleStep_; }
ensureSizeIsEnough(2, maxSize, CV_32FC4, outBuf);
void setAngleThresh(int angleThresh) { angleThresh_ = angleThresh; }
int getAngleThresh() const { return angleThresh_; }
posCount = Ballard_PosRotation_findPosInHist_gpu(hist, rows, cols, angleRange, outBuf.ptr<float4>(0), outBuf.ptr<int3>(1), maxSize, (float)minAngle, (float)angleStep, (float)dp, votesThreshold);
}
void setMinScale(double minScale) { minScale_ = minScale; }
double getMinScale() const { return minScale_; }
/////////////////////////////////////////
// POSITION & SCALE & ROTATION
void setMaxScale(double maxScale) { maxScale_ = maxScale; }
double getMaxScale() const { return maxScale_; }
double toRad(double a)
{
return a * CV_PI / 180.0;
}
void setScaleStep(double scaleStep) { scaleStep_ = scaleStep; }
double getScaleStep() const { return scaleStep_; }
double clampAngle(double a)
{
double res = a;
while (res > 360.0)
res -= 360.0;
while (res < 0)
res += 360.0;
void setScaleThresh(int scaleThresh) { scaleThresh_ = scaleThresh; }
int getScaleThresh() const { return scaleThresh_; }
return res;
}
void setPosThresh(int posThresh) { posThresh_ = posThresh; }
int getPosThresh() const { return posThresh_; }
bool angleEq(double a, double b, double eps = 1.0)
{
return (fabs(clampAngle(a - b)) <= eps);
}
private:
void processTempl();
void processImage();
class GHT_Guil_Full : public GHT_Pos
{
public:
AlgorithmInfo* info() const;
double xi_;
int levels_;
double angleEpsilon_;
GHT_Guil_Full();
double minAngle_;
double maxAngle_;
double angleStep_;
int angleThresh_;
protected:
void releaseImpl();
double minScale_;
double maxScale_;
double scaleStep_;
int scaleThresh_;
void processTempl();
void processImage();
int posThresh_;
struct Feature
{
......@@ -858,7 +658,6 @@ namespace
int maxSize;
void create(int levels, int maxCapacity, bool isTempl);
void release();
};
typedef void (*set_func_t)(PtrStepb p1_pos, PtrStepb p1_theta, PtrStepb p2_pos, PtrStepb d12, PtrStepb r1, PtrStepb r2);
......@@ -874,167 +673,126 @@ namespace
void calcScale(double angle);
void calcPosition(double angle, int angleVotes, double scale, int scaleVotes);
double xi;
int levels;
double angleEpsilon;
Feature templFeatures_;
Feature imageFeatures_;
double minAngle;
double maxAngle;
double angleStep;
int angleThresh;
std::vector< std::pair<double, int> > angles_;
std::vector< std::pair<double, int> > scales_;
double minScale;
double maxScale;
double scaleStep;
int scaleThresh;
GpuMat hist_;
std::vector<int> h_buf_;
};
double dp;
int posThresh;
double toRad(double a)
{
return a * CV_PI / 180.0;
}
Feature templFeatures;
Feature imageFeatures;
double clampAngle(double a)
{
double res = a;
std::vector< std::pair<double, int> > angles;
std::vector< std::pair<double, int> > scales;
while (res > 360.0)
res -= 360.0;
while (res < 0)
res += 360.0;
GpuMat hist;
std::vector<int> h_buf;
};
return res;
}
CV_INIT_ALGORITHM(GHT_Guil_Full, "GeneralizedHough_GPU.POSITION_SCALE_ROTATION",
obj.info()->addParam(obj, "minDist", obj.minDist, false, 0, 0,
"Minimum distance between the centers of the detected objects.");
obj.info()->addParam(obj, "maxSize", obj.maxSize, false, 0, 0,
"Maximal size of inner buffers.");
obj.info()->addParam(obj, "xi", obj.xi, false, 0, 0,
"Angle difference in degrees between two points in feature.");
obj.info()->addParam(obj, "levels", obj.levels, false, 0, 0,
"Feature table levels.");
obj.info()->addParam(obj, "angleEpsilon", obj.angleEpsilon, false, 0, 0,
"Maximal difference between angles that treated as equal.");
obj.info()->addParam(obj, "minAngle", obj.minAngle, false, 0, 0,
"Minimal rotation angle to detect in degrees.");
obj.info()->addParam(obj, "maxAngle", obj.maxAngle, false, 0, 0,
"Maximal rotation angle to detect in degrees.");
obj.info()->addParam(obj, "angleStep", obj.angleStep, false, 0, 0,
"Angle step in degrees.");
obj.info()->addParam(obj, "angleThresh", obj.angleThresh, false, 0, 0,
"Angle threshold.");
obj.info()->addParam(obj, "minScale", obj.minScale, false, 0, 0,
"Minimal scale to detect.");
obj.info()->addParam(obj, "maxScale", obj.maxScale, false, 0, 0,
"Maximal scale to detect.");
obj.info()->addParam(obj, "scaleStep", obj.scaleStep, false, 0, 0,
"Scale step.");
obj.info()->addParam(obj, "scaleThresh", obj.scaleThresh, false, 0, 0,
"Scale threshold.");
obj.info()->addParam(obj, "dp", obj.dp, false, 0, 0,
"Inverse ratio of the accumulator resolution to the image resolution.");
obj.info()->addParam(obj, "posThresh", obj.posThresh, false, 0, 0,
"Position threshold."));
GHT_Guil_Full::GHT_Guil_Full()
bool angleEq(double a, double b, double eps = 1.0)
{
maxSize = 1000;
xi = 90.0;
levels = 360;
angleEpsilon = 1.0;
minAngle = 0.0;
maxAngle = 360.0;
angleStep = 1.0;
angleThresh = 15000;
minScale = 0.5;
maxScale = 2.0;
scaleStep = 0.05;
scaleThresh = 1000;
dp = 1.0;
posThresh = 100;
return (fabs(clampAngle(a - b)) <= eps);
}
void GHT_Guil_Full::releaseImpl()
GeneralizedHoughGuilImpl::GeneralizedHoughGuilImpl()
{
GHT_Pos::releaseImpl();
maxBufferSize_ = 1000;
xi_ = 90.0;
levels_ = 360;
angleEpsilon_ = 1.0;
templFeatures.release();
imageFeatures.release();
minAngle_ = 0.0;
maxAngle_ = 360.0;
angleStep_ = 1.0;
angleThresh_ = 15000;
releaseVector(angles);
releaseVector(scales);
minScale_ = 0.5;
maxScale_ = 2.0;
scaleStep_ = 0.05;
scaleThresh_ = 1000;
hist.release();
releaseVector(h_buf);
posThresh_ = 100;
}
void GHT_Guil_Full::processTempl()
void GeneralizedHoughGuilImpl::processTempl()
{
using namespace cv::gpu::cudev::ght;
buildFeatureList(templEdges, templDx, templDy, templFeatures,
buildFeatureList(templEdges_, templDx_, templDy_, templFeatures_,
Guil_Full_setTemplFeatures, Guil_Full_buildTemplFeatureList_gpu,
true, templCenter);
true, templCenter_);
h_buf.resize(templFeatures.sizes.cols);
cudaSafeCall( cudaMemcpy(&h_buf[0], templFeatures.sizes.data, h_buf.size() * sizeof(int), cudaMemcpyDeviceToHost) );
templFeatures.maxSize = *max_element(h_buf.begin(), h_buf.end());
h_buf_.resize(templFeatures_.sizes.cols);
cudaSafeCall( cudaMemcpy(&h_buf_[0], templFeatures_.sizes.data, h_buf_.size() * sizeof(int), cudaMemcpyDeviceToHost) );
templFeatures_.maxSize = *std::max_element(h_buf_.begin(), h_buf_.end());
}
void GHT_Guil_Full::processImage()
void GeneralizedHoughGuilImpl::processImage()
{
using namespace cv::gpu::cudev::ght;
CV_Assert(levels > 0);
CV_Assert(templFeatures.sizes.cols == levels + 1);
CV_Assert(minAngle >= 0.0 && minAngle < maxAngle && maxAngle <= 360.0);
CV_Assert(angleStep > 0.0 && angleStep < 360.0);
CV_Assert(angleThresh > 0);
CV_Assert(minScale > 0.0 && minScale < maxScale);
CV_Assert(scaleStep > 0.0);
CV_Assert(scaleThresh > 0);
CV_Assert(dp > 0.0);
CV_Assert(posThresh > 0);
CV_Assert( levels_ > 0 );
CV_Assert( templFeatures_.sizes.cols == levels_ + 1 );
CV_Assert( minAngle_ >= 0.0 && minAngle_ < maxAngle_ && maxAngle_ <= 360.0 );
CV_Assert( angleStep_ > 0.0 && angleStep_ < 360.0 );
CV_Assert( angleThresh_ > 0 );
CV_Assert( minScale_ > 0.0 && minScale_ < maxScale_ );
CV_Assert( scaleStep_ > 0.0 );
CV_Assert( scaleThresh_ > 0 );
CV_Assert( dp_ > 0.0 );
CV_Assert( posThresh_ > 0 );
const double iAngleStep = 1.0 / angleStep;
const int angleRange = cvCeil((maxAngle - minAngle) * iAngleStep);
const double iAngleStep = 1.0 / angleStep_;
const int angleRange = cvCeil((maxAngle_ - minAngle_) * iAngleStep);
const double iScaleStep = 1.0 / scaleStep;
const int scaleRange = cvCeil((maxScale - minScale) * iScaleStep);
const double iScaleStep = 1.0 / scaleStep_;
const int scaleRange = cvCeil((maxScale_ - minScale_) * iScaleStep);
const double idp = 1.0 / dp;
const int histRows = cvCeil(imageSize.height * idp);
const int histCols = cvCeil(imageSize.width * idp);
const double idp = 1.0 / dp_;
const int histRows = cvCeil(imageSize_.height * idp);
const int histCols = cvCeil(imageSize_.width * idp);
ensureSizeIsEnough(histRows + 2, std::max(angleRange + 1, std::max(scaleRange + 1, histCols + 2)), CV_32SC1, hist);
h_buf.resize(std::max(angleRange + 1, scaleRange + 1));
ensureSizeIsEnough(histRows + 2, std::max(angleRange + 1, std::max(scaleRange + 1, histCols + 2)), CV_32SC1, hist_);
h_buf_.resize(std::max(angleRange + 1, scaleRange + 1));
ensureSizeIsEnough(2, maxSize, CV_32FC4, outBuf);
ensureSizeIsEnough(2, maxBufferSize_, CV_32FC4, outBuf_);
buildFeatureList(imageEdges, imageDx, imageDy, imageFeatures,
buildFeatureList(imageEdges_, imageDx_, imageDy_, imageFeatures_,
Guil_Full_setImageFeatures, Guil_Full_buildImageFeatureList_gpu,
false);
calcOrientation();
for (size_t i = 0; i < angles.size(); ++i)
for (size_t i = 0; i < angles_.size(); ++i)
{
const double angle = angles[i].first;
const int angleVotes = angles[i].second;
const double angle = angles_[i].first;
const int angleVotes = angles_[i].second;
calcScale(angle);
for (size_t j = 0; j < scales.size(); ++j)
for (size_t j = 0; j < scales_.size(); ++j)
{
const double scale = scales[j].first;
const int scaleVotes = scales[j].second;
const double scale = scales_[j].first;
const int scaleVotes = scales_[j].second;
calcPosition(angle, angleVotes, scale, scaleVotes);
}
}
}
void GHT_Guil_Full::Feature::create(int levels, int maxCapacity, bool isTempl)
void GeneralizedHoughGuilImpl::Feature::create(int levels, int maxCapacity, bool isTempl)
{
if (!isTempl)
{
......@@ -1058,128 +816,91 @@ namespace
maxSize = 0;
}
void GHT_Guil_Full::Feature::release()
void GeneralizedHoughGuilImpl::buildFeatureList(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy, Feature& features,
set_func_t set_func, build_func_t build_func, bool isTempl, Point2d center)
{
p1_pos.release();
p1_theta.release();
p2_pos.release();
CV_Assert( levels_ > 0 );
d12.release();
const double maxDist = sqrt((double) templSize_.width * templSize_.width + templSize_.height * templSize_.height) * maxScale_;
r1.release();
r2.release();
sizes.release();
maxSize = 0;
}
void GHT_Guil_Full::buildFeatureList(const GpuMat& edges, const GpuMat& dx, const GpuMat& dy, Feature& features,
set_func_t set_func, build_func_t build_func, bool isTempl, Point2d center)
{
CV_Assert(levels > 0);
const double maxDist = sqrt((double) templSize.width * templSize.width + templSize.height * templSize.height) * maxScale;
features.create(levels, maxSize, isTempl);
features.create(levels_, maxBufferSize_, isTempl);
set_func(features.p1_pos, features.p1_theta, features.p2_pos, features.d12, features.r1, features.r2);
buildEdgePointList(edges, dx, dy);
if (edgePointList.cols > 0)
if (edgePointList_.cols > 0)
{
build_func(edgePointList.ptr<unsigned int>(0), edgePointList.ptr<float>(1), edgePointList.cols,
features.sizes.ptr<int>(), maxSize, (float)xi, (float)angleEpsilon, levels, make_float2((float)center.x, (float)center.y), (float)maxDist);
build_func(edgePointList_.ptr<unsigned int>(0), edgePointList_.ptr<float>(1), edgePointList_.cols,
features.sizes.ptr<int>(), maxBufferSize_, (float)xi_, (float)angleEpsilon_, levels_, make_float2((float)center.x, (float)center.y), (float)maxDist);
}
}
void GHT_Guil_Full::calcOrientation()
void GeneralizedHoughGuilImpl::calcOrientation()
{
using namespace cv::gpu::cudev::ght;
const double iAngleStep = 1.0 / angleStep;
const int angleRange = cvCeil((maxAngle - minAngle) * iAngleStep);
const double iAngleStep = 1.0 / angleStep_;
const int angleRange = cvCeil((maxAngle_ - minAngle_) * iAngleStep);
hist.setTo(Scalar::all(0));
Guil_Full_calcOHist_gpu(templFeatures.sizes.ptr<int>(), imageFeatures.sizes.ptr<int>(0), hist.ptr<int>(),
(float)minAngle, (float)maxAngle, (float)angleStep, angleRange, levels, templFeatures.maxSize);
cudaSafeCall( cudaMemcpy(&h_buf[0], hist.data, h_buf.size() * sizeof(int), cudaMemcpyDeviceToHost) );
hist_.setTo(Scalar::all(0));
Guil_Full_calcOHist_gpu(templFeatures_.sizes.ptr<int>(), imageFeatures_.sizes.ptr<int>(0), hist_.ptr<int>(),
(float)minAngle_, (float)maxAngle_, (float)angleStep_, angleRange, levels_, templFeatures_.maxSize);
cudaSafeCall( cudaMemcpy(&h_buf_[0], hist_.data, h_buf_.size() * sizeof(int), cudaMemcpyDeviceToHost) );
angles.clear();
angles_.clear();
for (int n = 0; n < angleRange; ++n)
{
if (h_buf[n] >= angleThresh)
if (h_buf_[n] >= angleThresh_)
{
const double angle = minAngle + n * angleStep;
angles.push_back(std::make_pair(angle, h_buf[n]));
const double angle = minAngle_ + n * angleStep_;
angles_.push_back(std::make_pair(angle, h_buf_[n]));
}
}
}
void GHT_Guil_Full::calcScale(double angle)
void GeneralizedHoughGuilImpl::calcScale(double angle)
{
using namespace cv::gpu::cudev::ght;
const double iScaleStep = 1.0 / scaleStep;
const int scaleRange = cvCeil((maxScale - minScale) * iScaleStep);
const double iScaleStep = 1.0 / scaleStep_;
const int scaleRange = cvCeil((maxScale_ - minScale_) * iScaleStep);
hist.setTo(Scalar::all(0));
Guil_Full_calcSHist_gpu(templFeatures.sizes.ptr<int>(), imageFeatures.sizes.ptr<int>(0), hist.ptr<int>(),
(float)angle, (float)angleEpsilon, (float)minScale, (float)maxScale,
(float)iScaleStep, scaleRange, levels, templFeatures.maxSize);
cudaSafeCall( cudaMemcpy(&h_buf[0], hist.data, h_buf.size() * sizeof(int), cudaMemcpyDeviceToHost) );
hist_.setTo(Scalar::all(0));
Guil_Full_calcSHist_gpu(templFeatures_.sizes.ptr<int>(), imageFeatures_.sizes.ptr<int>(0), hist_.ptr<int>(),
(float)angle, (float)angleEpsilon_, (float)minScale_, (float)maxScale_,
(float)iScaleStep, scaleRange, levels_, templFeatures_.maxSize);
cudaSafeCall( cudaMemcpy(&h_buf_[0], hist_.data, h_buf_.size() * sizeof(int), cudaMemcpyDeviceToHost) );
scales.clear();
scales_.clear();
for (int s = 0; s < scaleRange; ++s)
{
if (h_buf[s] >= scaleThresh)
if (h_buf_[s] >= scaleThresh_)
{
const double scale = minScale + s * scaleStep;
scales.push_back(std::make_pair(scale, h_buf[s]));
const double scale = minScale_ + s * scaleStep_;
scales_.push_back(std::make_pair(scale, h_buf_[s]));
}
}
}
void GHT_Guil_Full::calcPosition(double angle, int angleVotes, double scale, int scaleVotes)
void GeneralizedHoughGuilImpl::calcPosition(double angle, int angleVotes, double scale, int scaleVotes)
{
using namespace cv::gpu::cudev::ght;
hist.setTo(Scalar::all(0));
Guil_Full_calcPHist_gpu(templFeatures.sizes.ptr<int>(), imageFeatures.sizes.ptr<int>(0), hist,
(float)angle, (float)angleEpsilon, (float)scale, (float)dp, levels, templFeatures.maxSize);
hist_.setTo(Scalar::all(0));
Guil_Full_calcPHist_gpu(templFeatures_.sizes.ptr<int>(), imageFeatures_.sizes.ptr<int>(0), hist_,
(float)angle, (float)angleEpsilon_, (float)scale, (float)dp_, levels_, templFeatures_.maxSize);
posCount = Guil_Full_findPosInHist_gpu(hist, outBuf.ptr<float4>(0), outBuf.ptr<int3>(1),
posCount, maxSize, (float)angle, angleVotes,
(float)scale, scaleVotes, (float)dp, posThresh);
posCount_ = Guil_Full_findPosInHist_gpu(hist_, outBuf_.ptr<float4>(0), outBuf_.ptr<int3>(1),
posCount_, maxBufferSize_, (float)angle, angleVotes,
(float)scale, scaleVotes, (float)dp_, posThresh_);
}
}
Ptr<gpu::GeneralizedHough> cv::gpu::GeneralizedHough::create(int method)
Ptr<GeneralizedHoughGuil> cv::gpu::createGeneralizedHoughGuil()
{
switch (method)
{
case cv::GeneralizedHough::GHT_POSITION:
CV_Assert( !GHT_Ballard_Pos_info_auto.name().empty() );
return new GHT_Ballard_Pos();
case (cv::GeneralizedHough::GHT_POSITION | cv::GeneralizedHough::GHT_SCALE):
CV_Assert( !GHT_Ballard_PosScale_info_auto.name().empty() );
return new GHT_Ballard_PosScale();
case (cv::GeneralizedHough::GHT_POSITION | cv::GeneralizedHough::GHT_ROTATION):
CV_Assert( !GHT_Ballard_PosRotation_info_auto.name().empty() );
return new GHT_Ballard_PosRotation();
case (cv::GeneralizedHough::GHT_POSITION | cv::GeneralizedHough::GHT_SCALE | cv::GeneralizedHough::GHT_ROTATION):
CV_Assert( !GHT_Guil_Full_info_auto.name().empty() );
return new GHT_Guil_Full();
default:
CV_Error(Error::StsBadArg, "Unsupported method");
return Ptr<GeneralizedHough>();
}
return new GeneralizedHoughGuilImpl;
}
#endif /* !defined (HAVE_CUDA) */
modules/gpuimgproc/test/test_hough.cpp
View file @ 7b87d72d
......@@ -193,7 +193,7 @@ PARAM_TEST_CASE(GeneralizedHough, cv::gpu::DeviceInfo, UseRoi)
{
};
GPU_TEST_P(GeneralizedHough, POSITION)
GPU_TEST_P(GeneralizedHough, Ballard)
{
const cv::gpu::DeviceInfo devInfo = GET_PARAM(0);
cv::gpu::setDevice(devInfo.deviceID());
......@@ -218,16 +218,16 @@ GPU_TEST_P(GeneralizedHough, POSITION)
templ.copyTo(imageROI);
}
cv::Ptr<cv::gpu::GeneralizedHough> hough = cv::gpu::GeneralizedHough::create(cv::GeneralizedHough::GHT_POSITION);
hough->set("votesThreshold", 200);
cv::Ptr<cv::GeneralizedHoughBallard> alg = cv::gpu::createGeneralizedHoughBallard();
alg->setVotesThreshold(200);
hough->setTemplate(loadMat(templ, useRoi));
alg->setTemplate(loadMat(templ, useRoi));
cv::gpu::GpuMat d_pos;
hough->detect(loadMat(image, useRoi), d_pos);
alg->detect(loadMat(image, useRoi), d_pos);
std::vector<cv::Vec4f> pos;
hough->downloadResults(d_pos, pos);
d_pos.download(pos);
ASSERT_EQ(gold_count, pos.size());
......
modules/imgproc/include/opencv2/imgproc.hpp
View file @ 7b87d72d
......@@ -688,39 +688,104 @@ public:
//! finds arbitrary template in the grayscale image using Generalized Hough Transform
class CV_EXPORTS GeneralizedHough : public Algorithm
{
public:
//! set template to search
virtual void setTemplate(InputArray templ, Point templCenter = Point(-1, -1)) = 0;
virtual void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter = Point(-1, -1)) = 0;
//! find template on image
virtual void detect(InputArray image, OutputArray positions, OutputArray votes = noArray()) = 0;
virtual void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes = noArray()) = 0;
//! Canny low threshold.
virtual void setCannyLowThresh(int cannyLowThresh) = 0;
virtual int getCannyLowThresh() const = 0;
//! Canny high threshold.
virtual void setCannyHighThresh(int cannyHighThresh) = 0;
virtual int getCannyHighThresh() const = 0;
//! Minimum distance between the centers of the detected objects.
virtual void setMinDist(double minDist) = 0;
virtual double getMinDist() const = 0;
//! Inverse ratio of the accumulator resolution to the image resolution.
virtual void setDp(double dp) = 0;
virtual double getDp() const = 0;
//! Maximal size of inner buffers.
virtual void setMaxBufferSize(int maxBufferSize) = 0;
virtual int getMaxBufferSize() const = 0;
};
//! Ballard, D.H. (1981). Generalizing the Hough transform to detect arbitrary shapes. Pattern Recognition 13 (2): 111-122.
//! Detects position only without traslation and rotation
class CV_EXPORTS GeneralizedHoughBallard : public GeneralizedHough
{
public:
//! R-Table levels.
virtual void setLevels(int levels) = 0;
virtual int getLevels() const = 0;
//! The accumulator threshold for the template centers at the detection stage. The smaller it is, the more false positions may be detected.
virtual void setVotesThreshold(int votesThreshold) = 0;
virtual int getVotesThreshold() const = 0;
};
//! Guil, N., González-Linares, J.M. and Zapata, E.L. (1999). Bidimensional shape detection using an invariant approach. Pattern Recognition 32 (6): 1025-1038.
class CV_EXPORTS GeneralizedHough : public Algorithm
//! Detects position, traslation and rotation
class CV_EXPORTS GeneralizedHoughGuil : public GeneralizedHough
{
public:
enum { GHT_POSITION = 0,
GHT_SCALE = 1,
GHT_ROTATION = 2
};
//! Angle difference in degrees between two points in feature.
virtual void setXi(double xi) = 0;
virtual double getXi() const = 0;
static Ptr<GeneralizedHough> create(int method);
//! Feature table levels.
virtual void setLevels(int levels) = 0;
virtual int getLevels() const = 0;
virtual ~GeneralizedHough();
//! Maximal difference between angles that treated as equal.
virtual void setAngleEpsilon(double angleEpsilon) = 0;
virtual double getAngleEpsilon() const = 0;
//! set template to search
void setTemplate(InputArray templ, int cannyThreshold = 100, Point templCenter = Point(-1, -1));
void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter = Point(-1, -1));
//! Minimal rotation angle to detect in degrees.
virtual void setMinAngle(double minAngle) = 0;
virtual double getMinAngle() const = 0;
//! find template on image
void detect(InputArray image, OutputArray positions, OutputArray votes = cv::noArray(), int cannyThreshold = 100);
void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes = cv::noArray());
//! Maximal rotation angle to detect in degrees.
virtual void setMaxAngle(double maxAngle) = 0;
virtual double getMaxAngle() const = 0;
void release();
//! Angle step in degrees.
virtual void setAngleStep(double angleStep) = 0;
virtual double getAngleStep() const = 0;
protected:
virtual void setTemplateImpl(const Mat& edges, const Mat& dx, const Mat& dy, Point templCenter) = 0;
virtual void detectImpl(const Mat& edges, const Mat& dx, const Mat& dy, OutputArray positions, OutputArray votes) = 0;
virtual void releaseImpl() = 0;
private:
Mat edges_;
Mat dx_;
Mat dy_;
//! Angle votes threshold.
virtual void setAngleThresh(int angleThresh) = 0;
virtual int getAngleThresh() const = 0;
//! Minimal scale to detect.
virtual void setMinScale(double minScale) = 0;
virtual double getMinScale() const = 0;
//! Maximal scale to detect.
virtual void setMaxScale(double maxScale) = 0;
virtual double getMaxScale() const = 0;
//! Scale step.
virtual void setScaleStep(double scaleStep) = 0;
virtual double getScaleStep() const = 0;
//! Scale votes threshold.
virtual void setScaleThresh(int scaleThresh) = 0;
virtual int getScaleThresh() const = 0;
//! Position votes threshold.
virtual void setPosThresh(int posThresh) = 0;
virtual int getPosThresh() const = 0;
};
......@@ -1355,6 +1420,14 @@ CV_EXPORTS_W double pointPolygonTest( InputArray contour, Point2f pt, bool measu
CV_EXPORTS Ptr<CLAHE> createCLAHE(double clipLimit = 40.0, Size tileGridSize = Size(8, 8));
//! Ballard, D.H. (1981). Generalizing the Hough transform to detect arbitrary shapes. Pattern Recognition 13 (2): 111-122.
//! Detects position only without traslation and rotation
CV_EXPORTS Ptr<GeneralizedHoughBallard> createGeneralizedHoughBallard();
//! Guil, N., González-Linares, J.M. and Zapata, E.L. (1999). Bidimensional shape detection using an invariant approach. Pattern Recognition 32 (6): 1025-1038.
//! Detects position, traslation and rotation
CV_EXPORTS Ptr<GeneralizedHoughGuil> createGeneralizedHoughGuil();
} // cv
#endif
modules/imgproc/src/generalized_hough.cpp
View file @ 7b87d72d
......@@ -45,17 +45,10 @@
using namespace cv;
// common
namespace
{
/////////////////////////////////////
// Common
template <typename T, class A> void releaseVector(std::vector<T, A>& v)
{
std::vector<T, A> empty;
empty.swap(v);
}
double toRad(double a)
{
return a * CV_PI / 180.0;
......@@ -66,70 +59,112 @@ namespace
return fabs(v) > std::numeric_limits<float>::epsilon();
}
class GHT_Pos : public GeneralizedHough
class GeneralizedHoughBase
{
public:
GHT_Pos();
protected:
void setTemplateImpl(const Mat& edges, const Mat& dx, const Mat& dy, Point templCenter);
void detectImpl(const Mat& edges, const Mat& dx, const Mat& dy, OutputArray positions, OutputArray votes);
void releaseImpl();
GeneralizedHoughBase();
virtual ~GeneralizedHoughBase() {}
void setTemplateImpl(InputArray templ, Point templCenter);
void setTemplateImpl(InputArray edges, InputArray dx, InputArray dy, Point templCenter);
void detectImpl(InputArray image, OutputArray positions, OutputArray votes);
void detectImpl(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes);
virtual void processTempl() = 0;
virtual void processImage() = 0;
int cannyLowThresh_;
int cannyHighThresh_;
double minDist_;
double dp_;
Size templSize_;
Point templCenter_;
Mat templEdges_;
Mat templDx_;
Mat templDy_;
Size imageSize_;
Mat imageEdges_;
Mat imageDx_;
Mat imageDy_;
std::vector<Vec4f> posOutBuf_;
std::vector<Vec3i> voteOutBuf_;
private:
void calcEdges(InputArray src, Mat& edges, Mat& dx, Mat& dy);
void filterMinDist();
void convertTo(OutputArray positions, OutputArray votes);
};
double minDist;
GeneralizedHoughBase::GeneralizedHoughBase()
{
cannyLowThresh_ = 50;
cannyHighThresh_ = 100;
minDist_ = 1.0;
dp_ = 1.0;
}
Size templSize;
Point templCenter;
Mat templEdges;
Mat templDx;
Mat templDy;
void GeneralizedHoughBase::calcEdges(InputArray _src, Mat& edges, Mat& dx, Mat& dy)
{
Mat src = _src.getMat();
Size imageSize;
Mat imageEdges;
Mat imageDx;
Mat imageDy;
CV_Assert( src.type() == CV_8UC1 );
CV_Assert( cannyLowThresh_ > 0 && cannyLowThresh_ < cannyHighThresh_ );
std::vector<Vec4f> posOutBuf;
std::vector<Vec3i> voteOutBuf;
};
Canny(src, edges, cannyLowThresh_, cannyHighThresh_);
Sobel(src, dx, CV_32F, 1, 0);
Sobel(src, dy, CV_32F, 0, 1);
}
GHT_Pos::GHT_Pos()
void GeneralizedHoughBase::setTemplateImpl(InputArray templ, Point templCenter)
{
minDist = 1.0;
calcEdges(templ, templEdges_, templDx_, templDy_);
if (templCenter == Point(-1, -1))
templCenter = Point(templEdges_.cols / 2, templEdges_.rows / 2);
templSize_ = templEdges_.size();
templCenter_ = templCenter;
processTempl();
}
void GHT_Pos::setTemplateImpl(const Mat& edges, const Mat& dx, const Mat& dy, Point templCenter_)
void GeneralizedHoughBase::setTemplateImpl(InputArray edges, InputArray dx, InputArray dy, Point templCenter)
{
templSize = edges.size();
templCenter = templCenter_;
edges.copyTo(templEdges);
dx.copyTo(templDx);
dy.copyTo(templDy);
edges.getMat().copyTo(templEdges_);
dx.getMat().copyTo(templDx_);
dy.getMat().copyTo(templDy_);
CV_Assert( templEdges_.type() == CV_8UC1 );
CV_Assert( templDx_.type() == CV_32FC1 && templDx_.size() == templEdges_.size() );
CV_Assert( templDy_.type() == templDx_.type() && templDy_.size() == templEdges_.size() );
if (templCenter == Point(-1, -1))
templCenter = Point(templEdges_.cols / 2, templEdges_.rows / 2);
templSize_ = templEdges_.size();
templCenter_ = templCenter;
processTempl();
}
void GHT_Pos::detectImpl(const Mat& edges, const Mat& dx, const Mat& dy, OutputArray positions, OutputArray votes)
void GeneralizedHoughBase::detectImpl(InputArray image, OutputArray positions, OutputArray votes)
{
imageSize = edges.size();
edges.copyTo(imageEdges);
dx.copyTo(imageDx);
dy.copyTo(imageDy);
calcEdges(image, imageEdges_, imageDx_, imageDy_);
imageSize_ = imageEdges_.size();
posOutBuf.clear();
voteOutBuf.clear();
posOutBuf_.clear();
voteOutBuf_.clear();
processImage();
if (!posOutBuf.empty())
if (!posOutBuf_.empty())
{
if (minDist > 1)
if (minDist_ > 1)
filterMinDist();
convertTo(positions, votes);
}
......@@ -141,21 +176,35 @@ namespace
}
}
void GHT_Pos::releaseImpl()
void GeneralizedHoughBase::detectImpl(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes)
{
templSize = Size();
templCenter = Point(-1, -1);
templEdges.release();
templDx.release();
templDy.release();
imageSize = Size();
imageEdges.release();
imageDx.release();
imageDy.release();
releaseVector(posOutBuf);
releaseVector(voteOutBuf);
edges.getMat().copyTo(imageEdges_);
dx.getMat().copyTo(imageDx_);
dy.getMat().copyTo(imageDy_);
CV_Assert( imageEdges_.type() == CV_8UC1 );
CV_Assert( imageDx_.type() == CV_32FC1 && imageDx_.size() == imageEdges_.size() );
CV_Assert( imageDy_.type() == imageDx_.type() && imageDy_.size() == imageEdges_.size() );
imageSize_ = imageEdges_.size();
posOutBuf_.clear();
voteOutBuf_.clear();
processImage();
if (!posOutBuf_.empty())
{
if (minDist_ > 1)
filterMinDist();
convertTo(positions, votes);
}
else
{
positions.release();
if (votes.needed())
votes.release();
}
}
class Vec3iGreaterThanIdx
......@@ -166,31 +215,31 @@ namespace
const Vec3i* arr;
};
void GHT_Pos::filterMinDist()
void GeneralizedHoughBase::filterMinDist()
{
size_t oldSize = posOutBuf.size();
const bool hasVotes = !voteOutBuf.empty();
size_t oldSize = posOutBuf_.size();
const bool hasVotes = !voteOutBuf_.empty();
CV_Assert(!hasVotes || voteOutBuf.size() == oldSize);
CV_Assert( !hasVotes || voteOutBuf_.size() == oldSize );
std::vector<Vec4f> oldPosBuf(posOutBuf);
std::vector<Vec3i> oldVoteBuf(voteOutBuf);
std::vector<Vec4f> oldPosBuf(posOutBuf_);
std::vector<Vec3i> oldVoteBuf(voteOutBuf_);
std::vector<size_t> indexies(oldSize);
for (size_t i = 0; i < oldSize; ++i)
indexies[i] = i;
std::sort(indexies.begin(), indexies.end(), Vec3iGreaterThanIdx(&oldVoteBuf[0]));
posOutBuf.clear();
voteOutBuf.clear();
posOutBuf_.clear();
voteOutBuf_.clear();
const int cellSize = cvRound(minDist);
const int gridWidth = (imageSize.width + cellSize - 1) / cellSize;
const int gridHeight = (imageSize.height + cellSize - 1) / cellSize;
const int cellSize = cvRound(minDist_);
const int gridWidth = (imageSize_.width + cellSize - 1) / cellSize;
const int gridHeight = (imageSize_.height + cellSize - 1) / cellSize;
std::vector< std::vector<Point2f> > grid(gridWidth * gridHeight);
const double minDist2 = minDist * minDist;
const double minDist2 = minDist_ * minDist_;
for (size_t i = 0; i < oldSize; ++i)
{
......@@ -239,108 +288,112 @@ namespace
{
grid[yCell * gridWidth + xCell].push_back(p);
posOutBuf.push_back(oldPosBuf[ind]);
posOutBuf_.push_back(oldPosBuf[ind]);
if (hasVotes)
voteOutBuf.push_back(oldVoteBuf[ind]);
voteOutBuf_.push_back(oldVoteBuf[ind]);
}
}
}
void GHT_Pos::convertTo(OutputArray _positions, OutputArray _votes)
void GeneralizedHoughBase::convertTo(OutputArray _positions, OutputArray _votes)
{
const int total = static_cast<int>(posOutBuf.size());
const bool hasVotes = !voteOutBuf.empty();
const int total = static_cast<int>(posOutBuf_.size());
const bool hasVotes = !voteOutBuf_.empty();
CV_Assert(!hasVotes || voteOutBuf.size() == posOutBuf.size());
CV_Assert( !hasVotes || voteOutBuf_.size() == posOutBuf_.size() );
_positions.create(1, total, CV_32FC4);
Mat positions = _positions.getMat();
Mat(1, total, CV_32FC4, &posOutBuf[0]).copyTo(positions);
Mat(1, total, CV_32FC4, &posOutBuf_[0]).copyTo(positions);
if (_votes.needed())
{
if (!hasVotes)
{
_votes.release();
}
else
{
_votes.create(1, total, CV_32SC3);
Mat votes = _votes.getMat();
Mat(1, total, CV_32SC3, &voteOutBuf[0]).copyTo(votes);
Mat(1, total, CV_32SC3, &voteOutBuf_[0]).copyTo(votes);
}
}
}
}
/////////////////////////////////////
// POSITION Ballard
// GeneralizedHoughBallard
class GHT_Ballard_Pos : public GHT_Pos
namespace
{
class GeneralizedHoughBallardImpl : public GeneralizedHoughBallard, private GeneralizedHoughBase
{
public:
AlgorithmInfo* info() const;
GeneralizedHoughBallardImpl();
GHT_Ballard_Pos();
void setTemplate(InputArray templ, Point templCenter) { setTemplateImpl(templ, templCenter); }
void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter) { setTemplateImpl(edges, dx, dy, templCenter); }
protected:
void releaseImpl();
void detect(InputArray image, OutputArray positions, OutputArray votes) { detectImpl(image, positions, votes); }
void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes) { detectImpl(edges, dx, dy, positions, votes); }
void setCannyLowThresh(int cannyLowThresh) { cannyLowThresh_ = cannyLowThresh; }
int getCannyLowThresh() const { return cannyLowThresh_; }
void setCannyHighThresh(int cannyHighThresh) { cannyHighThresh_ = cannyHighThresh; }
int getCannyHighThresh() const { return cannyHighThresh_; }
void setMinDist(double minDist) { minDist_ = minDist; }
double getMinDist() const { return minDist_; }
void setDp(double dp) { dp_ = dp; }
double getDp() const { return dp_; }
void setMaxBufferSize(int) { }
int getMaxBufferSize() const { return 0; }
void setLevels(int levels) { levels_ = levels; }
int getLevels() const { return levels_; }
void setVotesThreshold(int votesThreshold) { votesThreshold_ = votesThreshold; }
int getVotesThreshold() const { return votesThreshold_; }
private:
void processTempl();
void processImage();
virtual void calcHist();
virtual void findPosInHist();
void calcHist();
void findPosInHist();
int levels;
int votesThreshold;
double dp;
int levels_;
int votesThreshold_;
std::vector< std::vector<Point> > r_table;
Mat hist;
std::vector< std::vector<Point> > r_table_;
Mat hist_;
};
CV_INIT_ALGORITHM(GHT_Ballard_Pos, "GeneralizedHough.POSITION",
obj.info()->addParam(obj, "minDist", obj.minDist, false, 0, 0,
"Minimum distance between the centers of the detected objects.");
obj.info()->addParam(obj, "levels", obj.levels, false, 0, 0,
"R-Table levels.");
obj.info()->addParam(obj, "votesThreshold", obj.votesThreshold, false, 0, 0,
"The accumulator threshold for the template centers at the detection stage. The smaller it is, the more false positions may be detected.");
obj.info()->addParam(obj, "dp", obj.dp, false, 0, 0,
"Inverse ratio of the accumulator resolution to the image resolution."));
GHT_Ballard_Pos::GHT_Ballard_Pos()
GeneralizedHoughBallardImpl::GeneralizedHoughBallardImpl()
{
levels = 360;
votesThreshold = 100;
dp = 1.0;
levels_ = 360;
votesThreshold_ = 100;
}
void GHT_Ballard_Pos::releaseImpl()
void GeneralizedHoughBallardImpl::processTempl()
{
GHT_Pos::releaseImpl();
CV_Assert( levels_ > 0 );
releaseVector(r_table);
hist.release();
}
const double thetaScale = levels_ / 360.0;
void GHT_Ballard_Pos::processTempl()
{
CV_Assert(templEdges.type() == CV_8UC1);
CV_Assert(templDx.type() == CV_32FC1 && templDx.size() == templSize);
CV_Assert(templDy.type() == templDx.type() && templDy.size() == templSize);
CV_Assert(levels > 0);
r_table_.resize(levels_ + 1);
std::for_each(r_table_.begin(), r_table_.end(), std::mem_fun_ref(&std::vector<Point>::clear));
const double thetaScale = levels / 360.0;
r_table.resize(levels + 1);
for_each(r_table.begin(), r_table.end(), mem_fun_ref(&std::vector<Point>::clear));
for (int y = 0; y < templSize.height; ++y)
for (int y = 0; y < templSize_.height; ++y)
{
const uchar* edgesRow = templEdges.ptr(y);
const float* dxRow = templDx.ptr<float>(y);
const float* dyRow = templDy.ptr<float>(y);
const uchar* edgesRow = templEdges_.ptr(y);
const float* dxRow = templDx_.ptr<float>(y);
const float* dyRow = templDy_.ptr<float>(y);
for (int x = 0; x < templSize.width; ++x)
for (int x = 0; x < templSize_.width; ++x)
{
const Point p(x, y);
......@@ -348,42 +401,42 @@ namespace
{
const float theta = fastAtan2(dyRow[x], dxRow[x]);
const int n = cvRound(theta * thetaScale);
r_table[n].push_back(p - templCenter);
r_table_[n].push_back(p - templCenter_);
}
}
}
}
void GHT_Ballard_Pos::processImage()
void GeneralizedHoughBallardImpl::processImage()
{
calcHist();
findPosInHist();
}
void GHT_Ballard_Pos::calcHist()
void GeneralizedHoughBallardImpl::calcHist()
{
CV_Assert(imageEdges.type() == CV_8UC1);
CV_Assert(imageDx.type() == CV_32FC1 && imageDx.size() == imageSize);
CV_Assert(imageDy.type() == imageDx.type() && imageDy.size() == imageSize);
CV_Assert(levels > 0 && r_table.size() == static_cast<size_t>(levels + 1));
CV_Assert(dp > 0.0);
CV_Assert( imageEdges_.type() == CV_8UC1 );
CV_Assert( imageDx_.type() == CV_32FC1 && imageDx_.size() == imageSize_);
CV_Assert( imageDy_.type() == imageDx_.type() && imageDy_.size() == imageSize_);
CV_Assert( levels_ > 0 && r_table_.size() == static_cast<size_t>(levels_ + 1) );
CV_Assert( dp_ > 0.0 );
const double thetaScale = levels / 360.0;
const double idp = 1.0 / dp;
const double thetaScale = levels_ / 360.0;
const double idp = 1.0 / dp_;
hist.create(cvCeil(imageSize.height * idp) + 2, cvCeil(imageSize.width * idp) + 2, CV_32SC1);
hist.setTo(0);
hist_.create(cvCeil(imageSize_.height * idp) + 2, cvCeil(imageSize_.width * idp) + 2, CV_32SC1);
hist_.setTo(0);
const int rows = hist.rows - 2;
const int cols = hist.cols - 2;
const int rows = hist_.rows - 2;
const int cols = hist_.cols - 2;
for (int y = 0; y < imageSize.height; ++y)
for (int y = 0; y < imageSize_.height; ++y)
{
const uchar* edgesRow = imageEdges.ptr(y);
const float* dxRow = imageDx.ptr<float>(y);
const float* dyRow = imageDy.ptr<float>(y);
const uchar* edgesRow = imageEdges_.ptr(y);
const float* dxRow = imageDx_.ptr<float>(y);
const float* dyRow = imageDy_.ptr<float>(y);
for (int x = 0; x < imageSize.width; ++x)
for (int x = 0; x < imageSize_.width; ++x)
{
const Point p(x, y);
......@@ -392,7 +445,7 @@ namespace
const float theta = fastAtan2(dyRow[x], dxRow[x]);
const int n = cvRound(theta * thetaScale);
const std::vector<Point>& r_row = r_table[n];
const std::vector<Point>& r_row = r_table_[n];
for (size_t j = 0; j < r_row.size(); ++j)
{
......@@ -402,406 +455,131 @@ namespace
c.y = cvRound(c.y * idp);
if (c.x >= 0 && c.x < cols && c.y >= 0 && c.y < rows)
++hist.at<int>(c.y + 1, c.x + 1);
++hist_.at<int>(c.y + 1, c.x + 1);
}
}
}
}
}
void GHT_Ballard_Pos::findPosInHist()
void GeneralizedHoughBallardImpl::findPosInHist()
{
CV_Assert(votesThreshold > 0);
CV_Assert( votesThreshold_ > 0 );
const int histRows = hist.rows - 2;
const int histCols = hist.cols - 2;
const int histRows = hist_.rows - 2;
const int histCols = hist_.cols - 2;
for(int y = 0; y < histRows; ++y)
{
const int* prevRow = hist.ptr<int>(y);
const int* curRow = hist.ptr<int>(y + 1);
const int* nextRow = hist.ptr<int>(y + 2);
const int* prevRow = hist_.ptr<int>(y);
const int* curRow = hist_.ptr<int>(y + 1);
const int* nextRow = hist_.ptr<int>(y + 2);
for(int x = 0; x < histCols; ++x)
{
const int votes = curRow[x + 1];
if (votes > votesThreshold && votes > curRow[x] && votes >= curRow[x + 2] && votes > prevRow[x + 1] && votes >= nextRow[x + 1])
if (votes > votesThreshold_ && votes > curRow[x] && votes >= curRow[x + 2] && votes > prevRow[x + 1] && votes >= nextRow[x + 1])
{
posOutBuf.push_back(Vec4f(static_cast<float>(x * dp), static_cast<float>(y * dp), 1.0f, 0.0f));
voteOutBuf.push_back(Vec3i(votes, 0, 0));
posOutBuf_.push_back(Vec4f(static_cast<float>(x * dp_), static_cast<float>(y * dp_), 1.0f, 0.0f));
voteOutBuf_.push_back(Vec3i(votes, 0, 0));
}
}
}
}
}
/////////////////////////////////////
// POSITION & SCALE
class GHT_Ballard_PosScale : public GHT_Ballard_Pos
{
public:
AlgorithmInfo* info() const;
GHT_Ballard_PosScale();
protected:
void calcHist();
void findPosInHist();
double minScale;
double maxScale;
double scaleStep;
Ptr<GeneralizedHoughBallard> cv::createGeneralizedHoughBallard()
{
return new GeneralizedHoughBallardImpl;
}
class Worker;
friend class Worker;
};
// GeneralizedHoughGuil
CV_INIT_ALGORITHM(GHT_Ballard_PosScale, "GeneralizedHough.POSITION_SCALE",
obj.info()->addParam(obj, "minDist", obj.minDist, false, 0, 0,
"Minimum distance between the centers of the detected objects.");
obj.info()->addParam(obj, "levels", obj.levels, false, 0, 0,
"R-Table levels.");
obj.info()->addParam(obj, "votesThreshold", obj.votesThreshold, false, 0, 0,
"The accumulator threshold for the template centers at the detection stage. The smaller it is, the more false positions may be detected.");
obj.info()->addParam(obj, "dp", obj.dp, false, 0, 0,
"Inverse ratio of the accumulator resolution to the image resolution.");
obj.info()->addParam(obj, "minScale", obj.minScale, false, 0, 0,
"Minimal scale to detect.");
obj.info()->addParam(obj, "maxScale", obj.maxScale, false, 0, 0,
"Maximal scale to detect.");
obj.info()->addParam(obj, "scaleStep", obj.scaleStep, false, 0, 0,
"Scale step."));
GHT_Ballard_PosScale::GHT_Ballard_PosScale()
{
minScale = 0.5;
maxScale = 2.0;
scaleStep = 0.05;
}
class GHT_Ballard_PosScale::Worker : public ParallelLoopBody
namespace
{
class GeneralizedHoughGuilImpl : public GeneralizedHoughGuil, private GeneralizedHoughBase
{
public:
explicit Worker(GHT_Ballard_PosScale* base_) : base(base_) {}
GeneralizedHoughGuilImpl();
void operator ()(const Range& range) const;
void setTemplate(InputArray templ, Point templCenter) { setTemplateImpl(templ, templCenter); }
void setTemplate(InputArray edges, InputArray dx, InputArray dy, Point templCenter) { setTemplateImpl(edges, dx, dy, templCenter); }
private:
GHT_Ballard_PosScale* base;
};
void detect(InputArray image, OutputArray positions, OutputArray votes) { detectImpl(image, positions, votes); }
void detect(InputArray edges, InputArray dx, InputArray dy, OutputArray positions, OutputArray votes) { detectImpl(edges, dx, dy, positions, votes); }
void GHT_Ballard_PosScale::Worker::operator ()(const Range& range) const
{
const double thetaScale = base->levels / 360.0;
const double idp = 1.0 / base->dp;
void setCannyLowThresh(int cannyLowThresh) { cannyLowThresh_ = cannyLowThresh; }
int getCannyLowThresh() const { return cannyLowThresh_; }
for (int s = range.start; s < range.end; ++s)
{
const double scale = base->minScale + s * base->scaleStep;
void setCannyHighThresh(int cannyHighThresh) { cannyHighThresh_ = cannyHighThresh; }
int getCannyHighThresh() const { return cannyHighThresh_; }
Mat curHist(base->hist.size[1], base->hist.size[2], CV_32SC1, base->hist.ptr(s + 1), base->hist.step[1]);
void setMinDist(double minDist) { minDist_ = minDist; }
double getMinDist() const { return minDist_; }
for (int y = 0; y < base->imageSize.height; ++y)
{
const uchar* edgesRow = base->imageEdges.ptr(y);
const float* dxRow = base->imageDx.ptr<float>(y);
const float* dyRow = base->imageDy.ptr<float>(y);
void setDp(double dp) { dp_ = dp; }
double getDp() const { return dp_; }
for (int x = 0; x < base->imageSize.width; ++x)
{
const Point2d p(x, y);
void setMaxBufferSize(int maxBufferSize) { maxBufferSize_ = maxBufferSize; }
int getMaxBufferSize() const { return maxBufferSize_; }
if (edgesRow[x] && (notNull(dyRow[x]) || notNull(dxRow[x])))
{
const float theta = fastAtan2(dyRow[x], dxRow[x]);
const int n = cvRound(theta * thetaScale);
void setXi(double xi) { xi_ = xi; }
double getXi() const { return xi_; }
const std::vector<Point>& r_row = base->r_table[n];
void setLevels(int levels) { levels_ = levels; }
int getLevels() const { return levels_; }
for (size_t j = 0; j < r_row.size(); ++j)
{
Point2d d = r_row[j];
Point2d c = p - d * scale;
void setAngleEpsilon(double angleEpsilon) { angleEpsilon_ = angleEpsilon; }
double getAngleEpsilon() const { return angleEpsilon_; }
c.x *= idp;
c.y *= idp;
void setMinAngle(double minAngle) { minAngle_ = minAngle; }
double getMinAngle() const { return minAngle_; }
if (c.x >= 0 && c.x < base->hist.size[2] - 2 && c.y >= 0 && c.y < base->hist.size[1] - 2)
++curHist.at<int>(cvRound(c.y + 1), cvRound(c.x + 1));
}
}
}
}
}
}
void GHT_Ballard_PosScale::calcHist()
{
CV_Assert(imageEdges.type() == CV_8UC1);
CV_Assert(imageDx.type() == CV_32FC1 && imageDx.size() == imageSize);
CV_Assert(imageDy.type() == imageDx.type() && imageDy.size() == imageSize);
CV_Assert(levels > 0 && r_table.size() == static_cast<size_t>(levels + 1));
CV_Assert(dp > 0.0);
CV_Assert(minScale > 0.0 && minScale < maxScale);
CV_Assert(scaleStep > 0.0);
const double idp = 1.0 / dp;
const int scaleRange = cvCeil((maxScale - minScale) / scaleStep);
const int sizes[] = {scaleRange + 2, cvCeil(imageSize.height * idp) + 2, cvCeil(imageSize.width * idp) + 2};
hist.create(3, sizes, CV_32SC1);
hist.setTo(0);
parallel_for_(Range(0, scaleRange), Worker(this));
}
void GHT_Ballard_PosScale::findPosInHist()
{
CV_Assert(votesThreshold > 0);
const int scaleRange = hist.size[0] - 2;
const int histRows = hist.size[1] - 2;
const int histCols = hist.size[2] - 2;
for (int s = 0; s < scaleRange; ++s)
{
const float scale = static_cast<float>(minScale + s * scaleStep);
const Mat prevHist(histRows + 2, histCols + 2, CV_32SC1, hist.ptr(s), hist.step[1]);
const Mat curHist(histRows + 2, histCols + 2, CV_32SC1, hist.ptr(s + 1), hist.step[1]);
const Mat nextHist(histRows + 2, histCols + 2, CV_32SC1, hist.ptr(s + 2), hist.step[1]);
for(int y = 0; y < histRows; ++y)
{
const int* prevHistRow = prevHist.ptr<int>(y + 1);
const int* prevRow = curHist.ptr<int>(y);
const int* curRow = curHist.ptr<int>(y + 1);
const int* nextRow = curHist.ptr<int>(y + 2);
const int* nextHistRow = nextHist.ptr<int>(y + 1);
void setMaxAngle(double maxAngle) { maxAngle_ = maxAngle; }
double getMaxAngle() const { return maxAngle_; }
for(int x = 0; x < histCols; ++x)
{
const int votes = curRow[x + 1];
if (votes > votesThreshold &&
votes > curRow[x] &&
votes >= curRow[x + 2] &&
votes > prevRow[x + 1] &&
votes >= nextRow[x + 1] &&
votes > prevHistRow[x + 1] &&
votes >= nextHistRow[x + 1])
{
posOutBuf.push_back(Vec4f(static_cast<float>(x * dp), static_cast<float>(y * dp), scale, 0.0f));
voteOutBuf.push_back(Vec3i(votes, votes, 0));
}
}
}
}
}
void setAngleStep(double angleStep) { angleStep_ = angleStep; }
double getAngleStep() const { return angleStep_; }
/////////////////////////////////////
// POSITION & ROTATION
void setAngleThresh(int angleThresh) { angleThresh_ = angleThresh; }
int getAngleThresh() const { return angleThresh_; }
class GHT_Ballard_PosRotation : public GHT_Ballard_Pos
{
public:
AlgorithmInfo* info() const;
void setMinScale(double minScale) { minScale_ = minScale; }
double getMinScale() const { return minScale_; }
GHT_Ballard_PosRotation();
void setMaxScale(double maxScale) { maxScale_ = maxScale; }
double getMaxScale() const { return maxScale_; }
protected:
void calcHist();
void findPosInHist();
void setScaleStep(double scaleStep) { scaleStep_ = scaleStep; }
double getScaleStep() const { return scaleStep_; }
double minAngle;
double maxAngle;
double angleStep;
void setScaleThresh(int scaleThresh) { scaleThresh_ = scaleThresh; }
int getScaleThresh() const { return scaleThresh_; }
class Worker;
friend class Worker;
};
CV_INIT_ALGORITHM(GHT_Ballard_PosRotation, "GeneralizedHough.POSITION_ROTATION",
obj.info()->addParam(obj, "minDist", obj.minDist, false, 0, 0,
"Minimum distance between the centers of the detected objects.");
obj.info()->addParam(obj, "levels", obj.levels, false, 0, 0,
"R-Table levels.");
obj.info()->addParam(obj, "votesThreshold", obj.votesThreshold, false, 0, 0,
"The accumulator threshold for the template centers at the detection stage. The smaller it is, the more false positions may be detected.");
obj.info()->addParam(obj, "dp", obj.dp, false, 0, 0,
"Inverse ratio of the accumulator resolution to the image resolution.");
obj.info()->addParam(obj, "minAngle", obj.minAngle, false, 0, 0,
"Minimal rotation angle to detect in degrees.");
obj.info()->addParam(obj, "maxAngle", obj.maxAngle, false, 0, 0,
"Maximal rotation angle to detect in degrees.");
obj.info()->addParam(obj, "angleStep", obj.angleStep, false, 0, 0,
"Angle step in degrees."));
GHT_Ballard_PosRotation::GHT_Ballard_PosRotation()
{
minAngle = 0.0;
maxAngle = 360.0;
angleStep = 1.0;
}
class GHT_Ballard_PosRotation::Worker : public ParallelLoopBody
{
public:
explicit Worker(GHT_Ballard_PosRotation* base_) : base(base_) {}
void operator ()(const Range& range) const;
void setPosThresh(int posThresh) { posThresh_ = posThresh; }
int getPosThresh() const { return posThresh_; }
private:
GHT_Ballard_PosRotation* base;
};
void GHT_Ballard_PosRotation::Worker::operator ()(const Range& range) const
{
const double thetaScale = base->levels / 360.0;
const double idp = 1.0 / base->dp;
for (int a = range.start; a < range.end; ++a)
{
const double angle = base->minAngle + a * base->angleStep;
const double sinA = ::sin(toRad(angle));
const double cosA = ::cos(toRad(angle));
Mat curHist(base->hist.size[1], base->hist.size[2], CV_32SC1, base->hist.ptr(a + 1), base->hist.step[1]);
for (int y = 0; y < base->imageSize.height; ++y)
{
const uchar* edgesRow = base->imageEdges.ptr(y);
const float* dxRow = base->imageDx.ptr<float>(y);
const float* dyRow = base->imageDy.ptr<float>(y);
for (int x = 0; x < base->imageSize.width; ++x)
{
const Point2d p(x, y);
if (edgesRow[x] && (notNull(dyRow[x]) || notNull(dxRow[x])))
{
double theta = fastAtan2(dyRow[x], dxRow[x]) - angle;
if (theta < 0)
theta += 360.0;
const int n = cvRound(theta * thetaScale);
const std::vector<Point>& r_row = base->r_table[n];
for (size_t j = 0; j < r_row.size(); ++j)
{
Point2d d = r_row[j];
Point2d c = p - Point2d(d.x * cosA - d.y * sinA, d.x * sinA + d.y * cosA);
c.x *= idp;
c.y *= idp;
if (c.x >= 0 && c.x < base->hist.size[2] - 2 && c.y >= 0 && c.y < base->hist.size[1] - 2)
++curHist.at<int>(cvRound(c.y + 1), cvRound(c.x + 1));
}
}
}
}
}
}
void GHT_Ballard_PosRotation::calcHist()
{
CV_Assert(imageEdges.type() == CV_8UC1);
CV_Assert(imageDx.type() == CV_32FC1 && imageDx.size() == imageSize);
CV_Assert(imageDy.type() == imageDx.type() && imageDy.size() == imageSize);
CV_Assert(levels > 0 && r_table.size() == static_cast<size_t>(levels + 1));
CV_Assert(dp > 0.0);
CV_Assert(minAngle >= 0.0 && minAngle < maxAngle && maxAngle <= 360.0);
CV_Assert(angleStep > 0.0 && angleStep < 360.0);
const double idp = 1.0 / dp;
const int angleRange = cvCeil((maxAngle - minAngle) / angleStep);
const int sizes[] = {angleRange + 2, cvCeil(imageSize.height * idp) + 2, cvCeil(imageSize.width * idp) + 2};
hist.create(3, sizes, CV_32SC1);
hist.setTo(0);
parallel_for_(Range(0, angleRange), Worker(this));
}
void GHT_Ballard_PosRotation::findPosInHist()
{
CV_Assert(votesThreshold > 0);
const int angleRange = hist.size[0] - 2;
const int histRows = hist.size[1] - 2;
const int histCols = hist.size[2] - 2;
for (int a = 0; a < angleRange; ++a)
{
const float angle = static_cast<float>(minAngle + a * angleStep);
const Mat prevHist(histRows + 2, histCols + 2, CV_32SC1, hist.ptr(a), hist.step[1]);
const Mat curHist(histRows + 2, histCols + 2, CV_32SC1, hist.ptr(a + 1), hist.step[1]);
const Mat nextHist(histRows + 2, histCols + 2, CV_32SC1, hist.ptr(a + 2), hist.step[1]);
for(int y = 0; y < histRows; ++y)
{
const int* prevHistRow = prevHist.ptr<int>(y + 1);
const int* prevRow = curHist.ptr<int>(y);
const int* curRow = curHist.ptr<int>(y + 1);
const int* nextRow = curHist.ptr<int>(y + 2);
const int* nextHistRow = nextHist.ptr<int>(y + 1);
for(int x = 0; x < histCols; ++x)
{
const int votes = curRow[x + 1];
if (votes > votesThreshold &&
votes > curRow[x] &&
votes >= curRow[x + 2] &&
votes > prevRow[x + 1] &&
votes >= nextRow[x + 1] &&
votes > prevHistRow[x + 1] &&
votes >= nextHistRow[x + 1])
{
posOutBuf.push_back(Vec4f(static_cast<float>(x * dp), static_cast<float>(y * dp), 1.0f, angle));
voteOutBuf.push_back(Vec3i(votes, 0, votes));
}
}
}
}
}
/////////////////////////////////////////
// POSITION & SCALE & ROTATION
double clampAngle(double a)
{
double res = a;
while (res > 360.0)
res -= 360.0;
while (res < 0)
res += 360.0;
return res;
}
bool angleEq(double a, double b, double eps = 1.0)
{
return (fabs(clampAngle(a - b)) <= eps);
}
void processTempl();
void processImage();
class GHT_Guil_Full : public GHT_Pos
{
public:
AlgorithmInfo* info() const;
int maxBufferSize_;
double xi_;
int levels_;
double angleEpsilon_;
GHT_Guil_Full();
double minAngle_;
double maxAngle_;
double angleStep_;
int angleThresh_;
protected:
void releaseImpl();
double minScale_;
double maxScale_;
double scaleStep_;
int scaleThresh_;
void processTempl();
void processImage();
int posThresh_;
struct ContourPoint
{
......@@ -828,137 +606,92 @@ namespace
void calcScale(double angle);
void calcPosition(double angle, int angleVotes, double scale, int scaleVotes);
int maxSize;
double xi;
int levels;
double angleEpsilon;
std::vector< std::vector<Feature> > templFeatures_;
std::vector< std::vector<Feature> > imageFeatures_;
double minAngle;
double maxAngle;
double angleStep;
int angleThresh;
double minScale;
double maxScale;
double scaleStep;
int scaleThresh;
std::vector< std::pair<double, int> > angles_;
std::vector< std::pair<double, int> > scales_;
};
double dp;
int posThresh;
double clampAngle(double a)
{
double res = a;
std::vector< std::vector<Feature> > templFeatures;
std::vector< std::vector<Feature> > imageFeatures;
while (res > 360.0)
res -= 360.0;
while (res < 0)
res += 360.0;
std::vector< std::pair<double, int> > angles;
std::vector< std::pair<double, int> > scales;
};
return res;
}
CV_INIT_ALGORITHM(GHT_Guil_Full, "GeneralizedHough.POSITION_SCALE_ROTATION",
obj.info()->addParam(obj, "minDist", obj.minDist, false, 0, 0,
"Minimum distance between the centers of the detected objects.");
obj.info()->addParam(obj, "maxSize", obj.maxSize, false, 0, 0,
"Maximal size of inner buffers.");
obj.info()->addParam(obj, "xi", obj.xi, false, 0, 0,
"Angle difference in degrees between two points in feature.");
obj.info()->addParam(obj, "levels", obj.levels, false, 0, 0,
"Feature table levels.");
obj.info()->addParam(obj, "angleEpsilon", obj.angleEpsilon, false, 0, 0,
"Maximal difference between angles that treated as equal.");
obj.info()->addParam(obj, "minAngle", obj.minAngle, false, 0, 0,
"Minimal rotation angle to detect in degrees.");
obj.info()->addParam(obj, "maxAngle", obj.maxAngle, false, 0, 0,
"Maximal rotation angle to detect in degrees.");
obj.info()->addParam(obj, "angleStep", obj.angleStep, false, 0, 0,
"Angle step in degrees.");
obj.info()->addParam(obj, "angleThresh", obj.angleThresh, false, 0, 0,
"Angle threshold.");
obj.info()->addParam(obj, "minScale", obj.minScale, false, 0, 0,
"Minimal scale to detect.");
obj.info()->addParam(obj, "maxScale", obj.maxScale, false, 0, 0,
"Maximal scale to detect.");
obj.info()->addParam(obj, "scaleStep", obj.scaleStep, false, 0, 0,
"Scale step.");
obj.info()->addParam(obj, "scaleThresh", obj.scaleThresh, false, 0, 0,
"Scale threshold.");
obj.info()->addParam(obj, "dp", obj.dp, false, 0, 0,
"Inverse ratio of the accumulator resolution to the image resolution.");
obj.info()->addParam(obj, "posThresh", obj.posThresh, false, 0, 0,
"Position threshold."));
GHT_Guil_Full::GHT_Guil_Full()
bool angleEq(double a, double b, double eps = 1.0)
{
maxSize = 1000;
xi = 90.0;
levels = 360;
angleEpsilon = 1.0;
minAngle = 0.0;
maxAngle = 360.0;
angleStep = 1.0;
angleThresh = 15000;
minScale = 0.5;
maxScale = 2.0;
scaleStep = 0.05;
scaleThresh = 1000;
dp = 1.0;
posThresh = 100;
return (fabs(clampAngle(a - b)) <= eps);
}
void GHT_Guil_Full::releaseImpl()
GeneralizedHoughGuilImpl::GeneralizedHoughGuilImpl()
{
GHT_Pos::releaseImpl();
maxBufferSize_ = 1000;
xi_ = 90.0;
levels_ = 360;
angleEpsilon_ = 1.0;
releaseVector(templFeatures);
releaseVector(imageFeatures);
minAngle_ = 0.0;
maxAngle_ = 360.0;
angleStep_ = 1.0;
angleThresh_ = 15000;
releaseVector(angles);
releaseVector(scales);
minScale_ = 0.5;
maxScale_ = 2.0;
scaleStep_ = 0.05;
scaleThresh_ = 1000;
posThresh_ = 100;
}
void GHT_Guil_Full::processTempl()
void GeneralizedHoughGuilImpl::processTempl()
{
buildFeatureList(templEdges, templDx, templDy, templFeatures, templCenter);
buildFeatureList(templEdges_, templDx_, templDy_, templFeatures_, templCenter_);
}
void GHT_Guil_Full::processImage()
void GeneralizedHoughGuilImpl::processImage()
{
buildFeatureList(imageEdges, imageDx, imageDy, imageFeatures);
buildFeatureList(imageEdges_, imageDx_, imageDy_, imageFeatures_);
calcOrientation();
for (size_t i = 0; i < angles.size(); ++i)
for (size_t i = 0; i < angles_.size(); ++i)
{
const double angle = angles[i].first;
const int angleVotes = angles[i].second;
const double angle = angles_[i].first;
const int angleVotes = angles_[i].second;
calcScale(angle);
for (size_t j = 0; j < scales.size(); ++j)
for (size_t j = 0; j < scales_.size(); ++j)
{
const double scale = scales[j].first;
const int scaleVotes = scales[j].second;
const double scale = scales_[j].first;
const int scaleVotes = scales_[j].second;
calcPosition(angle, angleVotes, scale, scaleVotes);
}
}
}
void GHT_Guil_Full::buildFeatureList(const Mat& edges, const Mat& dx, const Mat& dy, std::vector< std::vector<Feature> >& features, Point2d center)
void GeneralizedHoughGuilImpl::buildFeatureList(const Mat& edges, const Mat& dx, const Mat& dy, std::vector< std::vector<Feature> >& features, Point2d center)
{
CV_Assert(levels > 0);
CV_Assert( levels_ > 0 );
const double maxDist = sqrt((double) templSize.width * templSize.width + templSize.height * templSize.height) * maxScale;
const double maxDist = sqrt((double) templSize_.width * templSize_.width + templSize_.height * templSize_.height) * maxScale_;
const double alphaScale = levels / 360.0;
const double alphaScale = levels_ / 360.0;
std::vector<ContourPoint> points;
getContourPoints(edges, dx, dy, points);
features.resize(levels + 1);
for_each(features.begin(), features.end(), mem_fun_ref(&std::vector<Feature>::clear));
for_each(features.begin(), features.end(), bind2nd(mem_fun_ref(&std::vector<Feature>::reserve), maxSize));
features.resize(levels_ + 1);
std::for_each(features.begin(), features.end(), std::mem_fun_ref(&std::vector<Feature>::clear));
std::for_each(features.begin(), features.end(), std::bind2nd(std::mem_fun_ref(&std::vector<Feature>::reserve), maxBufferSize_));
for (size_t i = 0; i < points.size(); ++i)
{
......@@ -968,7 +701,7 @@ namespace
{
ContourPoint p2 = points[j];
if (angleEq(p1.theta - p2.theta, xi, angleEpsilon))
if (angleEq(p1.theta - p2.theta, xi_, angleEpsilon_))
{
const Point2d d = p1.pos - p2.pos;
......@@ -988,18 +721,18 @@ namespace
const int n = cvRound(f.alpha12 * alphaScale);
if (features[n].size() < static_cast<size_t>(maxSize))
if (features[n].size() < static_cast<size_t>(maxBufferSize_))
features[n].push_back(f);
}
}
}
}
void GHT_Guil_Full::getContourPoints(const Mat& edges, const Mat& dx, const Mat& dy, std::vector<ContourPoint>& points)
void GeneralizedHoughGuilImpl::getContourPoints(const Mat& edges, const Mat& dx, const Mat& dy, std::vector<ContourPoint>& points)
{
CV_Assert(edges.type() == CV_8UC1);
CV_Assert(dx.type() == CV_32FC1 && dx.size == edges.size);
CV_Assert(dy.type() == dx.type() && dy.size == edges.size);
CV_Assert( edges.type() == CV_8UC1 );
CV_Assert( dx.type() == CV_32FC1 && dx.size == edges.size );
CV_Assert( dy.type() == dx.type() && dy.size == edges.size );
points.clear();
points.reserve(edges.size().area());
......@@ -1025,23 +758,23 @@ namespace
}
}
void GHT_Guil_Full::calcOrientation()
void GeneralizedHoughGuilImpl::calcOrientation()
{
CV_Assert(levels > 0);
CV_Assert(templFeatures.size() == static_cast<size_t>(levels + 1));
CV_Assert(imageFeatures.size() == templFeatures.size());
CV_Assert(minAngle >= 0.0 && minAngle < maxAngle && maxAngle <= 360.0);
CV_Assert(angleStep > 0.0 && angleStep < 360.0);
CV_Assert(angleThresh > 0);
CV_Assert( levels_ > 0 );
CV_Assert( templFeatures_.size() == static_cast<size_t>(levels_ + 1) );
CV_Assert( imageFeatures_.size() == templFeatures_.size() );
CV_Assert( minAngle_ >= 0.0 && minAngle_ < maxAngle_ && maxAngle_ <= 360.0 );
CV_Assert( angleStep_ > 0.0 && angleStep_ < 360.0 );
CV_Assert( angleThresh_ > 0 );
const double iAngleStep = 1.0 / angleStep;
const int angleRange = cvCeil((maxAngle - minAngle) * iAngleStep);
const double iAngleStep = 1.0 / angleStep_;
const int angleRange = cvCeil((maxAngle_ - minAngle_) * iAngleStep);
std::vector<int> OHist(angleRange + 1, 0);
for (int i = 0; i <= levels; ++i)
for (int i = 0; i <= levels_; ++i)
{
const std::vector<Feature>& templRow = templFeatures[i];
const std::vector<Feature>& imageRow = imageFeatures[i];
const std::vector<Feature>& templRow = templFeatures_[i];
const std::vector<Feature>& imageRow = imageFeatures_[i];
for (size_t j = 0; j < templRow.size(); ++j)
{
......@@ -1052,45 +785,45 @@ namespace
Feature imF = imageRow[k];
const double angle = clampAngle(imF.p1.theta - templF.p1.theta);
if (angle >= minAngle && angle <= maxAngle)
if (angle >= minAngle_ && angle <= maxAngle_)
{
const int n = cvRound((angle - minAngle) * iAngleStep);
const int n = cvRound((angle - minAngle_) * iAngleStep);
++OHist[n];
}
}
}
}
angles.clear();
angles_.clear();
for (int n = 0; n < angleRange; ++n)
{
if (OHist[n] >= angleThresh)
if (OHist[n] >= angleThresh_)
{
const double angle = minAngle + n * angleStep;
angles.push_back(std::make_pair(angle, OHist[n]));
const double angle = minAngle_ + n * angleStep_;
angles_.push_back(std::make_pair(angle, OHist[n]));
}
}
}
void GHT_Guil_Full::calcScale(double angle)
void GeneralizedHoughGuilImpl::calcScale(double angle)
{
CV_Assert(levels > 0);
CV_Assert(templFeatures.size() == static_cast<size_t>(levels + 1));
CV_Assert(imageFeatures.size() == templFeatures.size());
CV_Assert(minScale > 0.0 && minScale < maxScale);
CV_Assert(scaleStep > 0.0);
CV_Assert(scaleThresh > 0);
CV_Assert( levels_ > 0 );
CV_Assert( templFeatures_.size() == static_cast<size_t>(levels_ + 1) );
CV_Assert( imageFeatures_.size() == templFeatures_.size() );
CV_Assert( minScale_ > 0.0 && minScale_ < maxScale_ );
CV_Assert( scaleStep_ > 0.0 );
CV_Assert( scaleThresh_ > 0 );
const double iScaleStep = 1.0 / scaleStep;
const int scaleRange = cvCeil((maxScale - minScale) * iScaleStep);
const double iScaleStep = 1.0 / scaleStep_;
const int scaleRange = cvCeil((maxScale_ - minScale_) * iScaleStep);
std::vector<int> SHist(scaleRange + 1, 0);
for (int i = 0; i <= levels; ++i)
for (int i = 0; i <= levels_; ++i)
{
const std::vector<Feature>& templRow = templFeatures[i];
const std::vector<Feature>& imageRow = imageFeatures[i];
const std::vector<Feature>& templRow = templFeatures_[i];
const std::vector<Feature>& imageRow = imageFeatures_[i];
for (size_t j = 0; j < templRow.size(); ++j)
{
......@@ -1102,12 +835,12 @@ namespace
{
Feature imF = imageRow[k];
if (angleEq(imF.p1.theta, templF.p1.theta, angleEpsilon))
if (angleEq(imF.p1.theta, templF.p1.theta, angleEpsilon_))
{
const double scale = imF.d12 / templF.d12;
if (scale >= minScale && scale <= maxScale)
if (scale >= minScale_ && scale <= maxScale_)
{
const int s = cvRound((scale - minScale) * iScaleStep);
const int s = cvRound((scale - minScale_) * iScaleStep);
++SHist[s];
}
}
......@@ -1115,39 +848,39 @@ namespace
}
}
scales.clear();
scales_.clear();
for (int s = 0; s < scaleRange; ++s)
{
if (SHist[s] >= scaleThresh)
if (SHist[s] >= scaleThresh_)
{
const double scale = minScale + s * scaleStep;
scales.push_back(std::make_pair(scale, SHist[s]));
const double scale = minScale_ + s * scaleStep_;
scales_.push_back(std::make_pair(scale, SHist[s]));
}
}
}
void GHT_Guil_Full::calcPosition(double angle, int angleVotes, double scale, int scaleVotes)
void GeneralizedHoughGuilImpl::calcPosition(double angle, int angleVotes, double scale, int scaleVotes)
{
CV_Assert(levels > 0);
CV_Assert(templFeatures.size() == static_cast<size_t>(levels + 1));
CV_Assert(imageFeatures.size() == templFeatures.size());
CV_Assert(dp > 0.0);
CV_Assert(posThresh > 0);
CV_Assert( levels_ > 0 );
CV_Assert( templFeatures_.size() == static_cast<size_t>(levels_ + 1) );
CV_Assert( imageFeatures_.size() == templFeatures_.size() );
CV_Assert( dp_ > 0.0 );
CV_Assert( posThresh_ > 0 );
const double sinVal = sin(toRad(angle));
const double cosVal = cos(toRad(angle));
const double idp = 1.0 / dp;
const double idp = 1.0 / dp_;
const int histRows = cvCeil(imageSize.height * idp);
const int histCols = cvCeil(imageSize.width * idp);
const int histRows = cvCeil(imageSize_.height * idp);
const int histCols = cvCeil(imageSize_.width * idp);
Mat DHist(histRows + 2, histCols + 2, CV_32SC1, Scalar::all(0));
for (int i = 0; i <= levels; ++i)
for (int i = 0; i <= levels_; ++i)
{
const std::vector<Feature>& templRow = templFeatures[i];
const std::vector<Feature>& imageRow = imageFeatures[i];
const std::vector<Feature>& templRow = templFeatures_[i];
const std::vector<Feature>& imageRow = imageFeatures_[i];
for (size_t j = 0; j < templRow.size(); ++j)
{
......@@ -1165,7 +898,7 @@ namespace
{
Feature imF = imageRow[k];
if (angleEq(imF.p1.theta, templF.p1.theta, angleEpsilon))
if (angleEq(imF.p1.theta, templF.p1.theta, angleEpsilon_))
{
Point2d c1, c2;
......@@ -1195,101 +928,17 @@ namespace
{
const int votes = curRow[x + 1];
if (votes > posThresh && votes > curRow[x] && votes >= curRow[x + 2] && votes > prevRow[x + 1] && votes >= nextRow[x + 1])
if (votes > posThresh_ && votes > curRow[x] && votes >= curRow[x + 2] && votes > prevRow[x + 1] && votes >= nextRow[x + 1])
{
posOutBuf.push_back(Vec4f(static_cast<float>(x * dp), static_cast<float>(y * dp), static_cast<float>(scale), static_cast<float>(angle)));
voteOutBuf.push_back(Vec3i(votes, scaleVotes, angleVotes));
posOutBuf_.push_back(Vec4f(static_cast<float>(x * dp_), static_cast<float>(y * dp_), static_cast<float>(scale), static_cast<float>(angle)));
voteOutBuf_.push_back(Vec3i(votes, scaleVotes, angleVotes));
}
}
}
}
}
Ptr<GeneralizedHough> cv::GeneralizedHough::create(int method)
{
switch (method)
{
case GHT_POSITION:
CV_Assert( !GHT_Ballard_Pos_info_auto.name().empty() );
return new GHT_Ballard_Pos();
case (GHT_POSITION | GHT_SCALE):
CV_Assert( !GHT_Ballard_PosScale_info_auto.name().empty() );
return new GHT_Ballard_PosScale();
case (GHT_POSITION | GHT_ROTATION):
CV_Assert( !GHT_Ballard_PosRotation_info_auto.name().empty() );
return new GHT_Ballard_PosRotation();
case (GHT_POSITION | GHT_SCALE | GHT_ROTATION):
CV_Assert( !GHT_Guil_Full_info_auto.name().empty() );
return new GHT_Guil_Full();
}
CV_Error(CV_StsBadArg, "Unsupported method");
return Ptr<GeneralizedHough>();
}
cv::GeneralizedHough::~GeneralizedHough()
{
}
void cv::GeneralizedHough::setTemplate(InputArray _templ, int cannyThreshold, Point templCenter)
{
Mat templ = _templ.getMat();
CV_Assert(templ.type() == CV_8UC1);
CV_Assert(cannyThreshold > 0);
Canny(templ, edges_, cannyThreshold / 2, cannyThreshold);
Sobel(templ, dx_, CV_32F, 1, 0);
Sobel(templ, dy_, CV_32F, 0, 1);
if (templCenter == Point(-1, -1))
templCenter = Point(templ.cols / 2, templ.rows / 2);
setTemplateImpl(edges_, dx_, dy_, templCenter);
}
void cv::GeneralizedHough::setTemplate(InputArray _edges, InputArray _dx, InputArray _dy, Point templCenter)
{
Mat edges = _edges.getMat();
Mat dx = _dx.getMat();
Mat dy = _dy.getMat();
if (templCenter == Point(-1, -1))
templCenter = Point(edges.cols / 2, edges.rows / 2);
setTemplateImpl(edges, dx, dy, templCenter);
}
void cv::GeneralizedHough::detect(InputArray _image, OutputArray positions, OutputArray votes, int cannyThreshold)
{
Mat image = _image.getMat();
CV_Assert(image.type() == CV_8UC1);
CV_Assert(cannyThreshold > 0);
Canny(image, edges_, cannyThreshold / 2, cannyThreshold);
Sobel(image, dx_, CV_32F, 1, 0);
Sobel(image, dy_, CV_32F, 0, 1);
detectImpl(edges_, dx_, dy_, positions, votes);
}
void cv::GeneralizedHough::detect(InputArray _edges, InputArray _dx, InputArray _dy, OutputArray positions, OutputArray votes)
{
cv::Mat edges = _edges.getMat();
cv::Mat dx = _dx.getMat();
cv::Mat dy = _dy.getMat();
detectImpl(edges, dx, dy, positions, votes);
}
void cv::GeneralizedHough::release()
Ptr<GeneralizedHoughGuil> cv::createGeneralizedHoughGuil()
{
edges_.release();
dx_.release();
dy_.release();
releaseImpl();
return new GeneralizedHoughGuilImpl;
}
samples/gpu/generalized_hough.cpp
View file @ 7b87d72d
......@@ -5,13 +5,12 @@
#include "opencv2/core.hpp"
#include "opencv2/core/utility.hpp"
#include "opencv2/imgproc.hpp"
#include "opencv2/gpu.hpp"
#include "opencv2/gpuimgproc.hpp"
#include "opencv2/highgui.hpp"
#include "opencv2/contrib.hpp"
using namespace std;
using namespace cv;
using cv::gpu::GpuMat;
static Mat loadImage(const string& name)
{
......@@ -29,8 +28,7 @@ int main(int argc, const char* argv[])
CommandLineParser cmd(argc, argv,
"{ image i | pic1.png | input image }"
"{ template t | templ.png | template image }"
"{ scale s | | estimate scale }"
"{ rotation r | | estimate rotation }"
"{ full | | estimate scale and rotation }"
"{ gpu | | use gpu version }"
"{ minDist | 100 | minimum distance between the centers of the detected objects }"
"{ levels | 360 | R-Table levels }"
......@@ -45,7 +43,7 @@ int main(int argc, const char* argv[])
"{ minAngle | 0 | minimal rotation angle to detect in degrees }"
"{ maxAngle | 360 | maximal rotation angle to detect in degrees }"
"{ angleStep | 1 | angle step in degrees }"
"{ maxSize | 1000 | maximal size of inner buffers }"
"{ maxBufSize | 1000 | maximal size of inner buffers }"
"{ help h ? | | print help message }"
);
......@@ -59,8 +57,7 @@ int main(int argc, const char* argv[])
const string templName = cmd.get<string>("template");
const string imageName = cmd.get<string>("image");
const bool estimateScale = cmd.has("scale");
const bool estimateRotation = cmd.has("rotation");
const bool full = cmd.has("full");
const bool useGpu = cmd.has("gpu");
const double minDist = cmd.get<double>("minDist");
const int levels = cmd.get<int>("levels");
......@@ -75,7 +72,7 @@ int main(int argc, const char* argv[])
const double minAngle = cmd.get<double>("minAngle");
const double maxAngle = cmd.get<double>("maxAngle");
const double angleStep = cmd.get<double>("angleStep");
const int maxSize = cmd.get<int>("maxSize");
const int maxBufSize = cmd.get<int>("maxBufSize");
if (!cmd.check())
{
......@@ -86,93 +83,69 @@ int main(int argc, const char* argv[])
Mat templ = loadImage(templName);
Mat image = loadImage(imageName);
int method = cv::GeneralizedHough::GHT_POSITION;
if (estimateScale)
method += cv::GeneralizedHough::GHT_SCALE;
if (estimateRotation)
method += cv::GeneralizedHough::GHT_ROTATION;
Ptr<GeneralizedHough> alg;
if (!full)
{
Ptr<GeneralizedHoughBallard> ballard = useGpu ? gpu::createGeneralizedHoughBallard() : createGeneralizedHoughBallard();
ballard->setMinDist(minDist);
ballard->setLevels(levels);
ballard->setDp(dp);
ballard->setMaxBufferSize(maxBufSize);
ballard->setVotesThreshold(votesThreshold);
alg = ballard;
}
else
{
Ptr<GeneralizedHoughGuil> guil = useGpu ? gpu::createGeneralizedHoughGuil() : createGeneralizedHoughGuil();
guil->setMinDist(minDist);
guil->setLevels(levels);
guil->setDp(dp);
guil->setMaxBufferSize(maxBufSize);
guil->setMinAngle(minAngle);
guil->setMaxAngle(maxAngle);
guil->setAngleStep(angleStep);
guil->setAngleThresh(angleThresh);
guil->setMinScale(minScale);
guil->setMaxScale(maxScale);
guil->setScaleStep(scaleStep);
guil->setScaleThresh(scaleThresh);
guil->setPosThresh(posThresh);
alg = guil;
}
vector<Vec4f> position;
cv::TickMeter tm;
TickMeter tm;
if (useGpu)
{
GpuMat d_templ(templ);
GpuMat d_image(image);
GpuMat d_position;
Ptr<gpu::GeneralizedHough> d_hough = gpu::GeneralizedHough::create(method);
d_hough->set("minDist", minDist);
d_hough->set("levels", levels);
d_hough->set("dp", dp);
d_hough->set("maxSize", maxSize);
if (estimateScale && estimateRotation)
{
d_hough->set("angleThresh", angleThresh);
d_hough->set("scaleThresh", scaleThresh);
d_hough->set("posThresh", posThresh);
}
else
{
d_hough->set("votesThreshold", votesThreshold);
}
if (estimateScale)
{
d_hough->set("minScale", minScale);
d_hough->set("maxScale", maxScale);
d_hough->set("scaleStep", scaleStep);
}
if (estimateRotation)
{
d_hough->set("minAngle", minAngle);
d_hough->set("maxAngle", maxAngle);
d_hough->set("angleStep", angleStep);
}
d_hough->setTemplate(d_templ);
gpu::GpuMat d_templ(templ);
gpu::GpuMat d_image(image);
gpu::GpuMat d_position;
alg->setTemplate(d_templ);
tm.start();
d_hough->detect(d_image, d_position);
d_hough->downloadResults(d_position, position);
alg->detect(d_image, d_position);
d_position.download(position);
tm.stop();
}
else
{
Ptr<GeneralizedHough> hough = GeneralizedHough::create(method);
hough->set("minDist", minDist);
hough->set("levels", levels);
hough->set("dp", dp);
if (estimateScale && estimateRotation)
{
hough->set("angleThresh", angleThresh);
hough->set("scaleThresh", scaleThresh);
hough->set("posThresh", posThresh);
hough->set("maxSize", maxSize);
}
else
{
hough->set("votesThreshold", votesThreshold);
}
if (estimateScale)
{
hough->set("minScale", minScale);
hough->set("maxScale", maxScale);
hough->set("scaleStep", scaleStep);
}
if (estimateRotation)
{
hough->set("minAngle", minAngle);
hough->set("maxAngle", maxAngle);
hough->set("angleStep", angleStep);
}
hough->setTemplate(templ);
alg->setTemplate(templ);
tm.start();
hough->detect(image, position);
alg->detect(image, position);
tm.stop();
}
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment