Rachel Oliver, Air Force Institute of Technology; Michael Albert, University of Texas El Paso; Olac Fuentes, The University of Texas at El Paso; Dmitry Savransky, Cornell University
Keywords: Artificial Intelligence, Machine Learning, Space Domain Awareness, Object Tracking, Event-based Vision Sensors, Space-based
Abstract:
Building off of previously developed methods for grouping and classifying Event-based Vision Sensor (EVS) space domain awareness (SDA) event data, we demonstrate the customization and application of Random Sample Consensus (RANSAC) for grouping, and various machine learning (ML) model types including random forest, dense and convolutional neural networks (DNNs/CNNs) for group classification. While previous methods have removed event time information to apply classical two-dimensional image processing techniques, we retain the time information and develop three-dimensional native techniques to exploit it in order to maximize noise rejection, event association, and classification performance. We develop a methodology to allow off-the-shelf classifier models to use the 3-dimensional event information. Additionally we create these methods to process streamed event data, instead of static batch data of previous approaches. Finally, we generate synthetic data from a custom EVS simulator and evaluate the performance of the combined clustering and classifying algorithms. In their most effective configuration, the models provide .9827 true positive rate (TPR) and 0.9924 true negative rate (TNR) for satellite event grouping and classification on the validation data sets. The development of these techniques represents a significant step-forward in processing native EVS data, and demonstrates a leap forward in performance versus previous methods, moving EVS one step closer to operational SDA use.
Date of Conference: September 16-19, 2025
Track: Machine Learning for SDA Applications