I want to be able to take a photo of say, a circular piece of cardboard on the ground or something, and then paint some circles on it, then take a photo from the same angle and distance etc. OpenCV as I am using it will only correctly account for differences unless the photos are completely lined up the exact same place in the image. My goal is to either scan the image for something circular and be able to only run openCV on the circular parts of the image that has been cut out as a fix for this. Does anyone have any resources they could point me to?