Lukas Simkus
Research Mentor(s): Mojtaba Akhavan-Tafti
Mentor Department: Climate and Space Sciences and Engineering
Authors: Lukas Simkus, Mojtaba Akhavan-Tafti
Session: Session 1 (9:00am – 9:50am)
Presentation Type: Poster 26
Abstract
Space debris poses a significant threat to space missions, with over 100 million human-generated objects larger than 1 mm orbiting Earth. Centimeter- and millimeter-sized debris are the highest risk to space missions operating in Low Earth Orbit (LEO), as these pieces of small debris are capable of doing catastrophic damage at orbital velocities averaging 22,500 MPH. A 2021 National Science & Technology Council report states that less than 1% of mission-critical debris is tracked, and existing systems fail to characterize debris smaller than 10cm. Additionally, debris in GEO can persist for centuries due to negligible atmospheric drag, creating a growing field of debris made out of satellites and fragments. The issue is only growing, as in 2018, SpaceX received approval to launch 12,000 satellites, followed by a 2019 request for 30,000 more. This project addresses these challenges by developing a program to detect, track, and characterize small debris using images from the NEOWISE (Near-Earth Object Wide-field Infrared Survey Explorer) telescope. Space debris appears as linear streaks in these images due to their rapid motion across the telescope’s fixed field of view during long exposure times. To isolate streaks, the program uses median filtering, adaptive histogram equalization, Otsu thresholding, morphological operations, and Hough transforms, with celestial coordinates cross-referenced to known objects using astropy’s World Coordinate System. An initial implementation of the program used lenient thresholds to prioritize broad detection, identifying 319 potential streaks analyzing 11079 frames but achieving low precision (4.4%, 14 valid detections). The current implementation, analyzing 2,697 NEOWISE frames, used strict threshold filters and achieved 100% accuracy (5/5 detections valid) in streak detection and 100% accuracy (5/5 detections valid) in object correlation. While this method eliminates false positives, its strict parameters risk overlooking subtle or faint streaks, needing further validation across expanded datasets to ensure accuracy and generalize detection rates. The program also enables the discovery of previously unknown debris by detecting streaks uncorrelated with cataloged objects. This capability enhances the tracking of lethal non-trackable debris, supporting collision avoidance, mission planning, and the long-term sustainability of space operations. By working to monitor previously undetectable debris, this research project supports the safe operation of valuable space assets worldwide.
