_InterlockedExchangeAdd is a Visual Studio intrinsic that's available
for all architectures and in all VS versions that we care about. It's also
faster than the underscore-less function, since it's an intrinsic.
We also don't need to declare it ourselves.

It is, however, a Visual Studio-specific intrinsic, so I changed the
preprocessing condition accordingly.

Fixes <http://code.opencv.org/issues/3365>.

moved some headers to precomp.hpp

Fixed Viz compilation: removed CV_EXPORTS from cv::viz::Color. The macro forced instantiation of all Matx<double, 4, 1> constructors and triggered StaticAssert there.

Fixed compilation errors: removed using namespace from hpp file. This led to pulling cv::ACCESS_MASK to global namespace where conflict with the same name in 'windows.h'

changed Youtube link to our channel

This is something that should be left to the user, that's the whole
point of OpenCV_INCLUDE_DIRS.

Without limits included, several CUDA related files fail to compile with
GCC on Ubuntu:

modules/cudaimgproc/src/hough_lines.cpp:136:9: error: ‘numeric_limits’ is not a member of ‘std’

Conflicts:
	.gitignore
	modules/contrib/src/detection_based_tracker.cpp
	modules/core/include/opencv2/core/core.hpp
	modules/core/include/opencv2/core/internal.hpp
	modules/core/src/gpumat.cpp
	modules/core/src/opengl.cpp
	modules/gpu/src/cuda/safe_call.hpp
	modules/highgui/src/cap.cpp
	modules/imgproc/include/opencv2/imgproc/imgproc.hpp
	modules/ocl/doc/image_processing.rst
	modules/ocl/include/opencv2/ocl/ocl.hpp
	modules/ocl/perf/perf_haar.cpp
	modules/ocl/src/haar.cpp
	modules/ocl/src/imgproc.cpp
	modules/ocl/src/kmeans.cpp
	modules/ocl/src/svm.cpp
	modules/ocl/test/test_objdetect.cpp
	samples/ocl/adaptive_bilateral_filter.cpp

CMake's get_filename_component with NAME_WE parameters, interprets first period as start of extension. For. e.g file name with out extension of 'this.is.a.text.file.txt' will be taken as 'this'. Hence using NAME with regex replacement to get 'this.is.a.text.file'

In Python 3, version 3.3+, PyUnicode_AsUTF8() provides similar
functionality to Python 2's PyString_AsString().

In older versions of Python 3, there is no public function to provide
the same functionality.  However, the "internal" _PyUnicode_AsString()
does provide that functionality, so use it to replace
PyString_AsString().

With this patch, cv2 should compile for Python 3.[0-2].

Also, I made a separate macro for the current function name, which
helps simplify a lot of code that uses it.

Cleaned up adaptive bilateral filtering, added support for gaussian interpolation, updated sample and docs