Speedup about 30% on 6730M GPU.

1. Use macro defines for some parameters(radius).
2. Reduce local memory usage.
3. Fix accuracy problem on Intel GPU.

Add backwards portability for OpenCL 1.1 when OpenCV executables
are compiled with OpenCL 1.2 profile support.

A modification of the JavaCameraView is needed to avoid a crash when the
app is exited. It is a good practice to remove the callback after the
stopPreview method.

fixing precommit build on ubuntu-64 (now it takes OCL headers from non-default OCL SDK, but libs from default one)

The orientation of convexHull's result is currently the opposite of what the
documentation would suggest:

>>> import cv2, numpy as np
>>> points = np.array([[0,0],[0,1],[1,0]], dtype=np.int32)
>>> cv2.convexHull(points, clockwise=False)
array([[[1, 0]],
       [[0, 1]],
       [[0, 0]]], dtype=int32)
>>> cv2.convexHull(points, clockwise=True)
array([[[0, 0]],
       [[0, 1]],
       [[1, 0]]], dtype=int32)

Changing the function itself is probably not a good idea at this point, so
this fixes the documentation by flipping the coordinate system.

I also removed the mention of the origin, since it's irrelevant for this
function.

This also fixes the bug where the Sphinx version gets detected as
"Sphinx vX.Y.Z" instead of "X.Y.Z" on Unix-like systems.