Extended Target Tracking Algorithm Based on Morphological Matching Clustering in Near Spaced Environment

‍ ‍Aiming at the problems of low computational efficiency and inaccurate tracking of traditional extended target tracking （ETT） algorithm under the scenario where the targets are near spaced， a tracking algorithm combining the partition of morphological matching clustering measurement set and Gaussian inverse Wishart probability hypothesis density （GIW-PHD） filter is proposed in this paper. Firstly， the predicted target state is obtained by the GIW-PHD filter. Density based spatial clustering of applications with noise（DBSCAN） algorithm is used to preliminarily partition the measurement set. On this basis， the prediction component with higher weight is used to judge the mixed measurement cluster of multiple adjacent targets， and then the fuzzy c-means （FCM） algorithm with elliptical shape constraints is used to partition the mixed clusters to obtain more accurate partition results. Finally， the partition results are integrated and sent to the GIW-PHD filter to update the target state. The simulation results show that the proposed method can make full use of the predicted target position and shape information obtaining by GIW-PHD filter to accurately partition the mixed clusters， and thus to fulfill fast and accurate estimation of the kinematic state and extended state for the adjacent extended targets.