cuda_fast : multi stream safety (#10906)

* CUDA_Features2D/FAST Asynchronous test

* cuda_fast : multi stream safety

* Use parallel_for instead of OpenMP

* make tmpRow proper size before copyTo to avoid reallocated tmpCol

* do the transposition without creating temporary array; replace TAB with spaces.

* revert the previous commit

> cap_ffmpeg_impl.hpp:1710:11: error: 'uchar' does not name a type

about BYPRODUCTS bin/opencv-340.jar

include relu6 ocl kernel and layer fusion support
Signed-off-by: Li Peng <peng.li@intel.com>

* Use only absolute prior boxes explicit sizes. Remove scales attributes.

* Simplified PriorBox layer forward pass

For using 'OpenCV for Android SDK' with Android Studio projects.

To allow runtime disabling of AVX512-SKX via 'OPENCV_CPU_DISABLE' parameter

* Samples: reworked bgfg_bgsegm

* fixup! Samples: reworked bgfg_bgsegm

* add (untested) JavaCamera2View class

* initial fixes

* minor cleanup

* exclude JavaCamera2View from build for older SDK version

* fix method name typo

* add asserts + sanity checks

* fix preview format checks

* fix the memory leak

* export cvtTwoPlaneYUVtoBGR for Java usage

* add handling for two-plane YUV frames (C wrapper still missing)

* add two-plane cvtColor helper function

* fix warnings

* actually use the new cvtColorTwoPlane helper func

* fix wrong output matrix size

* fix wrong conversion type

* simplify method signature, add error condition

* minor fixes to Mat types

* remove leftover semaphore from camera api 1

* android: JavaCamera2View minor refactoring

- re-apply Java code style
- imports cleanup
- dump exceptions information

* android: JavaCamera2View: pause/resume fixes

* android: JavaCamera2View: fix mScale

* Make <array> #ifdef true for MSVC

I think MSVC had `std::array` for quite a while (possibly going back as far as VS 2012, but it's definitely there in 2015 and 2017. So I think `_MSC_VER` `1900` is a safe bet. Probably `1800` and maybe even `1700` could work as well but I can't test that locally.

* fix test