avoid `countNonZero()`, use `norm()`

GAPI (fluid): Sobel 3x3 optimization: CV_SIMD dynamic dispatching (#13133)

* GAPI (fluid): Sobel 3x3: remove template for run_sobel_row()

* GAPI (fluid): Sobel 3x3: dynamic dispatching of CV_SIMD code

* GAPI (fluid): Sobel 3x3 optimization: fixed CV_SIMD dynamic dispatcher

GAPI (fluid): optimization of Sobel 3x3 (#13070)

* GAPI: performance test for Sobel

* GAPI: performance test for Sobel w/FP32 input

* GAPI: Sobel speedup: 2.5x (U8) up to 10x (float)

* GAPI: Sobel 3x3 to support U8 into S16

* GAPI (fluid): Sobel 3x3 speedup: 10% (uchar), 1.5x (float)

* GAPI (fluid): Sobel 3x3 speedup: +10x (uchar), but -20% (float)

* GAPI (fluid): Sobel 3x3 speedup: +10% (float)

* GAPI (fluid): Sobel 3x3 speedup: +15% (float), +10% (uchar)

* GAPI (fluid): Sobel 3x3: address GCC warnings

* GAPI (fluid): Sobel 3x3: separate *.cpp file w/SIMD code

* GAPI (fluid): Sobel 3x3: fixed AVX2 code, AVX2 speedup 20-50% (uchar), 10-20% (float)

* GAPI (fluid): Sobel 3x3: fix CV_SIMD code for AVX2

* GAPI (fluid): Sobel 3x3: refactor

OpenCV 4.0.0-rc

avoid using of ENABLE_* flags

cap_libv4l depends on an external library (libv4l) yet is still larger
(1966 loc vs 1822 loc).
It was initially introduced copy pasting cap_v4l in order to offload
various color conversions to libv4l.
However nowadays we handle most of the needed color conversions inside
OpenCV. Our own implementation is better tested and (probably) also
better performing. (as it can optionally leverage IPP/ OpenCL)

Currently cap_v4l is better maintained and generally the code is in
better shape. There is however an API
difference in getting unconverted frames:
* on cap_libv4l one need to set `CV_CAP_MODE_GRAY=1` or
`CV_CAP_MODE_YUYV=1`
* on cap_v4l one needs to set `CV_CAP_PROP_CONVERT_RGB=0`

the latter is more flexible though as it also allows accessing undecoded
JPEG images.

fixes #4563

Fix #ifdef CV_SIMD128, should be #if CV_SIMD128.