杂波环境下基于meanshift的多目标数据关联

Data Association for Multiple Targets Based on MeanShift in Clutter

针对杂波背景下多目标的数据关联问题,提出了一种基于meanshift的多目标数据关联算法。该算法首次将meanshift用于多目标的数据关联问题中,首先对接收到的量测数据通过meanshift算法进行聚类处理,然后结合最近邻思想完成量测-航迹的关联,在目标密集或交叉的情况下,引入关联度的概念,与PDA相比,在提高关联精度的同时,也降低了计算量。蒙特卡罗仿真结果验证了该算法的可行性和有效性。

A new data association algorithm based on meanshifl for multiple targets in clutter was proposed. The meanshift method was applied firstly to data association for multiple targets. The proposed algorithm firstly classifies the received measurements, and then combines the nearest neighbor method to carry out the measurement-track association. By introducing the conception of association degree, association is exactly realized when targets are closed or crossing. Comparing with PDA algorithm, the precision of association can be improved while decreasing the computation burden. The Monte Carlo simulation results verify the feasibility and validity of the proposed algorithm.