Geosynchronous Satellite Detection and Tracking with WFOV Camera Arrays Using Spatio-Temporal Neural Networks (GEO-SPANN)
Abstract
Detection of low resolution, deep-space objects in wide field of view (WFOV) imaging systems can benefit from the emergence of temporally learned, appearance based detectors. The PANDORA sensor array, located in Maui at the Air Force Maui Optical and Supercomputing Site, is an exemplar of a scalable imaging architecture which can detect dim deep space objects while maintaining a WFOV. The PANDORA system captures 20x120 degree images of the night sky oriented along the GEO belt at a rate of two frames per minute. Prior work has established a baseline performance for the detection of Geosynchronous Earth Orbit (GEO) satellite objects using classical, feature based detectors, but has not leveraged the temporally rich data captured by PANDORA. This work extends the GEO object detection and tracking problem by implementing a spatio-temporal deep learning architecture (GEO-SPANN), further improving the state of the art in low resolution, deep-space object detection. Annotated sequential frames including object motion are used to train GEO-SPANN, which uses a two-stage CNN to provide a learned temporal mapping of GEO objects in sequences of annotated PANDORA images. We present the GEO object detection and tracking results of GEO-SPANN on sequences of 100 frames of PANDORA data. GEO-SPANN advances strategies for autonomous detection and tracking of GEO satellites, allowing PANDORA to be leveraged for orbit catalogue maintenance and space object anomaly detection.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 22, 2022
- Accession Number
- AD1200329
Entities
People
- Garrett Fitzgerald
- Ruixu Liu
- Vijayan Asari
Organizations
- Air Force Research Laboratory