Conversion was not ITU 601 conformant. Also add YVU420 format to allowed
zero copy access.

internally using ipp this conversion is now roughly 2x faster.
also add RGB/ BGR formats to allowed zero copy access.

possible source of memory leaks and unneeded complexity

by updating to c++ coding conventions

 - const correctness
 - use bool as return value
 - use explicit initialization instead of CLEAR macro/ memset
 - use cv namespace

the last kernel that allowed compiling this code was 2.6.37 which was
released almost 5 Years ago. So probably it does not get much testing
any more. Furthermore even back then one was better off using the V4L2
API.

The only change touching currently used code is the removal of the
global V4L2_SUPPORT variable.

Instead of chaining a bunch of sanity checks together with "&&", let's just have several asserts. That way, when an assert fails, you don't get a monsterous "<huge evil expression>
failed" error, but only the bit you care about, making your life rather a lot easier.

Conflicts:
	modules/imgcodecs/include/opencv2/imgcodecs.hpp

set GA_Update flag in order remove ERROR 8: band 1: Attempt to write to read only dataset inGDALRasterBand::Fill()

Current implementation returns NaN for some configs of FFMPEG, call default get_fps() instead (as also used by PROP_POS_MSEC)

build for some OCL drivers.
Some OCL vendors treat implicit scalar-vector conversions for
operators as errors when type conversion is required.