Easier to understand

fixed difference between SSE2 and C branches of StereoBM; they should now produce absolutely the same results

In the function cvInitIntrinsicParams2D the principal point for
normalized image coordinates is set to 0/0. This updates the function
to initialize the principal point at 0.5/0.5.

findHomagraphy can return empty Mat in master branch

fix thin-prism coefficients at 0's in the stereo calibration function unless the flag CALIB_THIN_PRISM_MODEL is explicitly set

Conflicts:
	modules/core/include/opencv2/core/cuda.hpp
	modules/cudacodec/src/thread.cpp
	modules/cudacodec/src/thread.hpp
	modules/superres/perf/perf_superres.cpp
	modules/superres/src/btv_l1_cuda.cpp
	modules/superres/src/optical_flow.cpp
	modules/videostab/src/global_motion.cpp
	modules/videostab/src/inpainting.cpp
	samples/cpp/stitching_detailed.cpp
	samples/cpp/videostab.cpp
	samples/gpu/stereo_multi.cpp

Looks like we don't actually use them; when creating precompiled headers
with Visual C++, a dummy .cpp is created at build time.

This should allow many algorithms to take advantage of more parallelization
technologies.