Using target radial length for data association in multiple-target tracking

Data association plays an important role in multi-target tracking. The traditional data association algorithm uses the nearest neighbor distance method. It will make mistake in larger echo density. Feature aided data association algorithm is tend of development in multi-target tracking. Incorporating target kinematics information and feature information can increase the information dimension, and enhance the association accuracy. In this work, feature aided nearest neighbor algorithm based on radial length of target is proposed. Comparing the traditional data association algorithm, the proposed algorithm can increase the times of correct association when targets move in parallel or move crosswise, and improve the performance of data association algorithm for multi-target tracking.