Space Object Detection in Images Using Matched Filter Bank and Bayesian Update

Abstract

Electrooptical sensors, when used to track space objects, are often used to produce detections of space objects for some orbit determination scheme. Instead, this paper proposes a series of methods to use electrooptical images directly in orbit determination. This work uses the signal-to-noise ratio optimal image filter, called a matched filter, to search for partially known space objects. By defining a metric for measuring matched filter template similarity, a bank of matched filters is efficiently defined by partitioning the prior knowledge set. Once partitioned sets are known, the matched filter bank can be localized to regions of the sky. A method for hypothesis testing the result of a matched filter for a space object is developed. Finally, a framework for orbit determination based on the matched filter result is developed. Simulation shows that the analytic results enable a better framework for implementing matched filters for low-signal-to-noise ratio object detection.