A computer vision process to detect and track space debris using ground-based optical telephoto images

Orbital debris travel at very high speeds in space. They pose a serious threat to the safety of both humans and machines in space and their number is ever increasing. Detecting and tracking orbital debris has been done primarily by radars in the past. A new approach is to detect and track orbital debris using optical telephoto images. This technique has the potential to detect debris of small sizes or of low radar reflectivities. This paper describes an approach which first extracts potential moving objects from two consecutive frames. Possible streak segments are then extracted from matching the two sets of potential moving objects in a 3-frame sequence. Lastly, consistent linear motions are extracted and reported over a longer observation period