Issue compiling OpenCV 4.6.0 on top of cuda 11.5

[cudawarped]

Please post the full output from the CMake configuration stage, see

'cv2.cuda.GpuMat' object has no attribute 'cols' when compiled from source

When I try to get cols and rows from CUDA GpuMat, pixel_number = gray.cols * gray.rows it results in error: AttributeError: 'cv2.cuda.GpuMat' object has no attribute 'cols' What is interesting is when I run this script using precompiled opencv-python & opencv-contrib-python 4.7.0, it never throws error. Only when I run it with OpenCV I compiled from source. Did I miss some options? My CMake command: CUDA_PATH="/usr/local/cuda-11.4" CFLAGS="-I/usr/local/cuda-11.4/include -I/opt/intel/compil…

for an example.