Any tips for reliable vision tracking of powercells? We’re looking to track them and running into issues with thresholding the entire cell without letting everything else in too.
We had good luck with hsl filtering, computing contours, filtering the contours to take a out the real small stuff. The ones remaining were turned into convex hulls. The hulls were checked for circularity and those that passed are power cells. This link provides some info on calculating circularity. https://www.learnopencv.com/blob-detection-using-opencv-python-c/