Tracklets Association Algorithm for Multitarget Tracking With Distributed Through-Wall Imaging Radar

In practical scenarios, obstacles, such as furniture or building structures, are commonly present within rooms, making multitarget tracking in distributed through-wall imaging radar (TWIR) a challenging task. Especially, the unobservable area caused by obstacles can result in the tracker generating multiple tracklets rather than a continuous and complete trajectory. Tracklets association is crucial for improving the readability of tracking results, but it remains an unsolved problem in distributed TWIR. In this article, we propose a tracklets association method for the continuity of target's identification (ID) in distributed TWIR. Specifically, first, the local tracklets can be get by mean-shift tracking method in each TWIR node. The global tracklets is composed of local tracklets after renumbering. Then, the similarity between global tracklets is calculated based on their motion characteristics. The tracklets-target's ID matching matrix (TTI-MM) is given to express the association relationship between global tracklets and target's ID, which can be solved by minimizing the difference between the similarity matrix and the TTI-MM. Based on the global tracklets marked with target's ID, the view-dependent features of the ghost is utilized for ghost ID. Finally, the performance of the proposed method is validated by simulation and experimental results.