Three-Dimensional Group Target Separation Detection Method Based on Ellipsoid Shape Reconstruction

An important challenge in group target tracking is the separation of individual targets within the group. Group target separation can lead to significant fluctuations in the position of the group center and the shape of the group, leading to decreased tracking accuracy and potential target loss. Moreover, when the target group consists of multiple objects with uncertain spatial positions, especially during separation, rapidly reconstructing the shape of the group target becomes challenging. This article proposes a novel method for detecting separation events and stable tracking to address related issues. Initially, we establish rapid ellipsoidal modeling of the group target shape through measurement mapping. Subsequently, group target separation events are predicted in real time by monitoring changes in ellipsoid volume between frames. Meanwhile, an adaptive association gate and a group clustering threshold are set to assist in separation assessment. In addition, utilize the pre-separation group target state to stabilize subgroups' tracking after separation. The simulation results demonstrate that the proposed algorithm effectively and timely detects group target separation and enhances the performance of tracking separated group targets.