Space-based tracking for long-distance multi-target in complex star background

Abstract

The space-based tracking for multi-target in complex star background is challenging due to the imaging interference of stars and the frequent target birth in field-of-view. In this paper, a multiple space debris tracking method using a dynamic threshold and an adaptive early-warning-boundary (EWB) is proposed. Firstly, the Gaussian expansion processing is used to extend the information of point targets and stars, and the grayscale likelihood function for the expanded candidate objects is calculated for state description. To reduce the interference arising from the stars around the targets, the dynamic threshold is taken as the criterion for distinguishing between stars and targets. To deal with the frequent targets birth, the EWB is generated using the motion characteristics. Furthermore, an identifying-and-tracking mechanism is established, which can identify the newborn targets on the EWB and estimate the state of targets inside of the EWB. Simulation results show that the multi-target tracking method is capable of reducing the rate of misjudgments and identifying newborn targets with low detection cost. Compared with the traditional tracking method using Gaussian mixture probability hypothesis density (GM-PHD), the proposed method can improve the estimation accuracy of targets’ number and state.