Event-Based Weak Target Detection and Tracking for Space Situational Awareness

Weak targets detection is a challenging task in space situational awareness. With the advantage of high dynamic range, dynamic vision sensors offer possible resolution for the issue. This paper utilizes an event-based camera to show a spatial imaging dataset and explores a Gaussian Mixture Model (GMM) clustering technique based on event data to address the detection of weak targets in the space. In the detection process, additional noise would be introduced, and the targets of interest may be faint and difficult to distinguish from sensor noise. This paper applies background filtering techniques for noise filtering and utilizes the GMM for detection. Experimental validation demonstrates that the combination of background filtering and GMM is particularly effective for detecting weak targets in space, exhibiting strong robustness. The GMM achieves an accuracy of 94.43% and a recall rate of 99.25% for weak targets. Overall, this paper discussed the challenges and prospects of faced by Event Based Camera (EBC) in Space Situational Awareness (SSA), and verify the feasibility and prospects of EBC for SSA.