I was wondering if someone knows how to take 2-D keypoints selected in raw images and map them to the orthomosaic image built from these raw images? I have a couple of thousand labels drawn on individual raw images (the label coordinates are relative to the raw image). Because raw images overlap, I have multiple labels over the same object in the ground when the object shows across multiple images. I have attached a couple of images where the labels are shown as labeled bounding boxes.
I think I can represent my labels as single 2-D keypoints in individual raw images. I imagine that once my orthomosaic image is created, I should be able to map these 2-D points into the orthomosaic in the same way 2-D image points are transformed into 3-D points. How can I do this using a python script? I'll appreciate any help. Thank you!