A Multi-target Tracking Evaluation Method Based on Fourier Transform Considering Shape Differences

Abstract

The invention relates to the technical field of multi-target tracking, and discloses a multi-target tracking evaluation method considering shape differences based on Fourier transform. The method is implemented by the following steps: Step 1: Obtain a real target state set and Estimating the target state set; Step 2, constructing the FOSPA distance between the real target state set and the estimated target state set; wherein, the FOSPA distance includes position error, potential error and shape error; Step 3, solve the FOSPA distance; Step 4 , and evaluate the performance of the tracking algorithm based on the obtained FOSPA distance, and give an evaluation conclusion. For the evaluation requirements of the example scene requiring consideration of the shape difference, the method proposed in the present invention can better meet the evaluation requirements, and the obtained evaluation conclusion is more intuitive and reasonable. In particular, when calculating the FOSPA distance for each tracking moment, it can achieve The algorithm performance is evaluated in the whole tracking process, so the method has good application value and promotion prospect.

Description

technical field [0001] The invention relates to the technical field of multi-target tracking, in particular to a multi-target tracking evaluation method based on Fourier transform considering shape differences. Background technique [0002] Multi-target tracking is one of the key functions of radar, electronic reconnaissance, air traffic control and other equipment / systems, and its tracking performance has a very important impact on the overall performance of the system. In recent years, the multi-target tracking algorithm based on random finite set (RFS) has become a research hotspot. This algorithm is different from the traditional multi-target tracking algorithm based on data association, and the traditional target tracking algorithm evaluation method cannot be used when evaluating its performance. . Therefore, in order to effectively evaluate the multi-object tracking algorithm based on RFS, Schuhmacher et al. have given the definition of optimal sub-pattern allocation ...

AI Technical Summary

Problems solved by technology

[0003] (1) Algorithms for some multi-target tracking scenarios, especially the tracking algorithms for group targets (a special form of multi-target activity, such as formation targets, cooperative combat targets, etc.) established by many scholars under the framework of RFS, multi-target The geometric shape between them is of great significance to the understanding of combat intent and threat estimation, but the above method does not include the measurement of the geometric shape difference, and cannot evaluate the estimation performance of the tracking algorithm for the shape, which is not conducive to researchers to choose a more suitable tracking algorithm;

[0004] (2) When comparing the performance of multiple tracking algorithms, if the target number estimation and target position estimation are equivalent, the OSPA distance values ​​in the above methods are very close, and the resolution is extremely poor, which is not conducive to researchers to obtain more effective tracking algorithms. The comparison result of

Images