Added workaround for Nvidia: take into account that 3-channel vector type takes 4*elem_size in local memory.

don't discard pure virtual methods in hdr_parser.py (was broken anyway)

cv::Matx could not be converted to c::Mat automatically

Based on commit feb74b12 from 2.4.9 branch.

fixed dependency of stitching module on xfeatures2d module as optional
fixed the initModule_xfeatures2d function that was called in module stitching since it is in another namespace than the standard cv one.