I am using a gstreamer pipeline with OpenCV VideoCapture on Jetson Nano. The source is captured in NV12 and I have to convert to BGR in the gstreamer pipeline. In order to optimize it, I thought I would use Graph-API. My understanding is that I would leave the pipeline in the NV12 format and then I would do something like the following for GAPI:
cv::GMat in;
cv::GMat out = cv::gapi::NV12toBGR(in);
cv::GComputation ac(in, out);
//videocapture stuff
ac.apply(input, output);
But as it turns out, NV12toBGR has two parameters for y and uv respectively. How are you supposed to use this function alongside a gstreamer pipeline that would capture in NV12?