aded singe kernel matchings

fixed warnings and error

replaced colls with strides

fixed warnings

replaced colls with strides everywhere

the number of images is estabished at run time

fixed braket warnings

replaced all enums with a single enum

removed whitespaces

fixed last issues

the matching class
this class contains the most important functions used in stereo correspondence
(StereoBM and StereoSGBM) call methods from these classes

fixed shadowing warning problems

included intrin.h header for popcnt instruction

replaced __popcnt with _mm_popcnt_u32() equivalent

fixed some warnings

fixed index warning

modules/stereo/src/stereo_binary_sgbm.cpp

@@ -47,7 +47,6 @@ by Heiko Hirschmuller.
 We match blocks rather than individual pixels, thus the algorithm is called
 SGBM (Semi-global block matching)
 */
-
 #include "precomp.hpp"
 #include <limits.h>
 
@@ -58,7 +57,6 @@ namespace cv
         typedef uchar PixType;
         typedef short CostType;
         typedef short DispType;
-
         enum { NR = 16, NR2 = NR/2 };
         struct StereoBinarySGBMParams
         {
@@ -74,7 +72,6 @@ namespace cv
                 speckleRange = 0;
                 mode = StereoBinarySGBM::MODE_SGBM;
             }
-
             StereoBinarySGBMParams( int _minDisparity, int _numDisparities, int _SADWindowSize,
                 int _P1, int _P2, int _disp12MaxDiff, int _preFilterCap,
                 int _uniquenessRatio, int _speckleWindowSize, int _speckleRange,
@@ -161,10 +158,8 @@ namespace cv
                 }
             }
             memset( cost, 0, width1*D*sizeof(cost[0]) );
-
             buffer -= minX2;
             cost -= minX1*D + minD; // simplify the cost indices inside the loop
-
 #if CV_SSE2
             volatile bool useSIMD = checkHardwareSupport(CV_CPU_SSE2);
 #endif
@@ -298,7 +293,6 @@ namespace cv
 
             volatile bool useSIMD = checkHardwareSupport(CV_CPU_SSE2);
 #endif
-
             const int ALIGN = 16;
             const int DISP_SHIFT = StereoMatcher::DISP_SHIFT;
             const int DISP_SCALE = (1 << DISP_SHIFT);
@@ -410,17 +404,14 @@ namespace cv
                                     for( d = 0; d < D; d++ )
                                         hsumAdd[d] = (CostType)(hsumAdd[d] + pixDiff[x + d]*scale);
                                 }
-
                                 if( y > 0 )
                                 {
                                     const CostType* hsumSub = hsumBuf + (std::max(y - SH2 - 1, 0) % hsumBufNRows)*costBufSize;
                                     const CostType* Cprev = !fullDP || y == 0 ? C : C - costBufSize;
-
                                     for( x = D; x < width1*D; x += D )
                                     {
                                         const CostType* pixAdd = pixDiff + std::min(x + SW2*D, (width1-1)*D);
                                         const CostType* pixSub = pixDiff + std::max(x - (SW2+1)*D, 0);
-
 #if CV_SSE2
                                         if( useSIMD )
                                         {
@@ -563,28 +554,21 @@ namespace cv
 #endif
                         {
                             int minL0 = MAX_COST, minL1 = MAX_COST, minL2 = MAX_COST, minL3 = MAX_COST;
-
                             for( d = 0; d < D; d++ )
                             {
                                 int Cpd = Cp[d], L0, L1, L2, L3;
-
                                 L0 = Cpd + std::min((int)Lr_p0[d], std::min(Lr_p0[d-1] + P1, std::min(Lr_p0[d+1] + P1, delta0))) - delta0;
                                 L1 = Cpd + std::min((int)Lr_p1[d], std::min(Lr_p1[d-1] + P1, std::min(Lr_p1[d+1] + P1, delta1))) - delta1;
                                 L2 = Cpd + std::min((int)Lr_p2[d], std::min(Lr_p2[d-1] + P1, std::min(Lr_p2[d+1] + P1, delta2))) - delta2;
                                 L3 = Cpd + std::min((int)Lr_p3[d], std::min(Lr_p3[d-1] + P1, std::min(Lr_p3[d+1] + P1, delta3))) - delta3;
-
                                 Lr_p[d] = (CostType)L0;
                                 minL0 = std::min(minL0, L0);
-
                                 Lr_p[d + D2] = (CostType)L1;
                                 minL1 = std::min(minL1, L1);
-
                                 Lr_p[d + D2*2] = (CostType)L2;
                                 minL2 = std::min(minL2, L2);
-
                                 Lr_p[d + D2*3] = (CostType)L3;
                                 minL3 = std::min(minL3, L3);
-
                                 Sp[d] = saturate_cast<CostType>(Sp[d] + L0 + L1 + L2 + L3);
                             }
                             minLr[0][xm] = (CostType)minL0;
@@ -593,7 +577,6 @@ namespace cv
                             minLr[0][xm+3] = (CostType)minL3;
                         }
                     }
-
                     if( pass == npasses )
                     {
                         for( x = 0; x < width; x++ )
@@ -610,21 +593,17 @@ namespace cv
                             if( npasses == 1 )
                             {
                                 int xm = x*NR2, xd = xm*D2;
-
                                 int minL0 = MAX_COST;
                                 int delta0 = minLr[0][xm + NR2] + P2;
                                 CostType* Lr_p0 = Lr[0] + xd + NRD2;
                                 Lr_p0[-1] = Lr_p0[D] = MAX_COST;
                                 CostType* Lr_p = Lr[0] + xd;
-
                                 const CostType* Cp = C + x*D;
-
 #if CV_SSE2
                                 if( useSIMD )
                                 {
                                     __m128i _P1 = _mm_set1_epi16((short)P1);
                                     __m128i _delta0 = _mm_set1_epi16((short)delta0);
-
                                     __m128i _minL0 = _mm_set1_epi16((short)minL0);
                                     __m128i _minS = _mm_set1_epi16(MAX_COST), _bestDisp = _mm_set1_epi16(-1);
                                     __m128i _d8 = _mm_setr_epi16(0, 1, 2, 3, 4, 5, 6, 7), _8 = _mm_set1_epi16(8);
@@ -645,7 +624,6 @@ namespace cv
                                         _bestDisp = _mm_xor_si128(_bestDisp, _mm_and_si128(_mm_xor_si128(_bestDisp,_d8), mask));
                                         _d8 = _mm_adds_epi16(_d8, _8);
                                     }
-
                                     short CV_DECL_ALIGNED(16) bestDispBuf[8];
                                     _mm_store_si128((__m128i*)bestDispBuf, _bestDisp);
                                     _minL0 = _mm_min_epi16(_minL0, _mm_srli_si128(_minL0, 8));
@@ -764,8 +742,7 @@ namespace cv
                 disparr.create( left.size(), CV_16S );
                 Mat disp = disparr.getMat();
                 computeDisparityBinarySGBM( left, right, disp, params, buffer );
-				medianBlur(disp, disp, 3);
-
+                // medianBlur(disp, disp, 3);
                 if( params.speckleWindowSize > 0 )
                     filterSpeckles(disp, (params.minDisparity - 1)*StereoMatcher::DISP_SCALE, params.speckleWindowSize,
                     StereoMatcher::DISP_SCALE*params.speckleRange, buffer);
@@ -818,7 +795,6 @@ namespace cv
                     << "P2" << params.P2
                     << "mode" << params.mode;
             }
-
             void read(const FileNode& fn)
             {
                 FileNode n = fn["name"];
@@ -835,14 +811,11 @@ namespace cv
                 params.P2 = (int)fn["P2"];
                 params.mode = (int)fn["mode"];
             }
-
             StereoBinarySGBMParams params;
             Mat buffer;
             static const char* name_;
         };
-
         const char* StereoBinarySGBMImpl::name_ = "StereoMatcher.SGBM";
-
         Ptr<StereoBinarySGBM> StereoBinarySGBM::create(int minDisparity, int numDisparities, int SADWindowSize,
             int P1, int P2, int disp12MaxDiff,
             int preFilterCap, int uniquenessRatio,
@@ -856,7 +829,6 @@ namespace cv
                 speckleWindowSize, speckleRange,
                 mode));
         }
-		
         typedef cv::Point_<short> Point2s;
     }
 }