考虑形状差异的RFS多目标跟踪性能评估方法  

Performance Evaluation Considering Shape Difference for Multi-target Tracking Based on Random Finite Set

作　　者：王碧垚 王永齐 顾鹏 WANG Bi-yao;WANG Yong-qi;GU Peng(The 29th Research Institute of China Electronics Technology Group Corporation,Chengdu 610036,China)

机构地区：[1]中国电子科技集团公司第二十九研究所,成都610036

出　　处：《火力与指挥控制》2021年第5期58-63,共6页Fire Control & Command Control

摘　　要：在对随机有限集(Random Finite Set,RFS)多目标跟踪算法进行性能评估时,最优子模式分配距离(Optimal Sub-Pattern Assignment,OSPA)及改进形式是最常用的指标。分析了采用OSPA距离的评估方法存在不完备的情况,考虑到多目标运动过程中组成的几何形状信息,通过引入形状差异度量,提出了一种改进的OSPA距离,并基于傅里叶描述子给出了形状差异度量的计算公式。仿真结果表明,针对部分RFS多目标跟踪性能评估场景,所提的改进方法可以得到更加完备的评估结论。While evaluating multi-target tracking algorithms based on random finite set(RFS),the original and improved optimal sub-pattern assignment(OSPA)distances are most frequently used indexes.After the shortage of original OSPA distance evaluation in some situations is analyzed and geometrical shape information formed in the multi-targets movement process is considered,a modified OSPA distance is proposed by introducing shape difference measurment.The formulae for shape difference measurement based on fourier descriptors are provided.Simulations demonstrate that in parts of RFS multi-targets tracking performance evaluation scenarios,the proposed modification method could obtain more complete evaluation conclusions.

关 键 词：多目标跟踪 性能评估 最优子模式分配距离 形状差异 

分 类 号：TJ0[兵器科学与技术—兵器发射理论与技术] TN953[电子电信—信号与信息处理]