Debris Classification in Low Earth Orbit

Space debris is an ever-growing problem in the field of astrophysics. A 2017 report from MIT tracked over 17,000 pieces of debris from broken satellites and space shuttles from high-powered telescopes and lasers–and that’s a lowball considering that there are reportedly over 500,000 pieces of debris that are smaller than a pebble in space, moving at over 17,500 miles per hour. Since this debris is primarily in a low-earth orbit where most space operations happen, such a large field of debris has proven more hazardous and difficult to avoid over time. We plan to help solve this problem by developing a computer vision neural network that can both detect the type and location of debris so that it can either perform evasive maneuvers or move in to capture the debris in an effort to stop the buildup of space junk. The first step will be to code a simulation of the orbit of a piece of debris around the Earth. Once this is done, we will be able to start training the computer vision to recognize the debris and, based on its classification, decide what course of action to take.