Hi,
I am working on a project which needs image processing pixel by pixel, I mean select pixels of certain color from image and get it’s coordinates but I am unable to do it with the documentation.
I am loading my images using OpenCV as cv::Mat.
Please help me achieve this as I am new in GPGPU.
Regards,
Hi,
You can use cv::cuda::GpuMat to get fast cv processing on gpu with threshold function / inRange function ocv::cuda implementation (see OpenCV CUDA - method that works like inRange() - OpenCV Q&A Forum),
Another option is using Nvidia Performance Primitives (NPP) , e.g. https://docs.nvidia.com/cuda/npp/group__image__threshold__operations.html
Best Regards
1 Like