I am using digital photogrammetry to build a point cloud of a river bank. The photos were taken from a moving raft, and I need to mask the raft and oar from the pictures. I used MATLAB to mask the raft, which was fairly easy, but I haven't been able to figure out how to mask the oar, which moves from picture to picture. There are 13,000+ pictures, so it needs to be a batch operation. I cannot just mask an area, because that masked area would cover a large portion of the picture. Does anyone have any tips on how to detect a moving object when the background is moving as well? I'd post pictures, but I don't have enough reputation.
Thanks!
This is a machine learning (ML) task. You can easily program primitive algorithms which would be able to solve your taks without any toolbox. If you do not have enough experience in math you may also look out for some existing ML toolboxes or algorithms which use ML to track objects, like this one:
http://personal.ee.surrey.ac.uk/Personal/Z.Kalal/tld.html