An enhanced VTS radar target tracking method addressing target splitting with integrated reasoning

The paper introduces a target-tracking technique designed to resolve the issue of target splitting in VTS radar imagery. The method consists of three parts: firstly, background subtraction is used for target detection in radar images, which results i n a large number of split targets affecting the tracking outcome; secondly, a target splitting reasoning mechanism based on the distance between targets and their orientation is designed to improve the detection results; finally, the optimized results a re tracked using Kalman filtering and the Hungarian algorithm. The experiments in the detection phase are divided into three groups: the first group does not use the reasoning mechanism and relies solely on background subtraction, the second group uses distance as the judgment factor, and the third group uses both distance and angle for judgment. The tracking phase is divided into two groups: the first tracks the detection results without the reasoning mechanism, and the second tracks the detection results using both distance and angle judgments. The experimental results show that the reasoning mechanism can significantly improve the effectiveness of target detection. Distance is the primary criterion for determining target splitti ng, whereas angle cannot be consistently relied upon as a basis for judgment. The effectiveness of this target tracking paradigm depends on the accuracy of detection, and the inclusion of the reasoning mechanism significantly enhances the final target tracking performance.