高分辨毫米波雷达与摄像融合目标跟踪方法

Target Tracking Method for Fusion of High-Resolution Millimeter-Wave Radar and Camera

为提高雷达感知对抗领域的雷达光电目标感知能力,针对高分辨毫米波雷达与摄像融合目标跟踪算法进行研究。针对单传感器性能局限与现有雷达光电融合跟踪鲁棒性较弱的问题,提出一种高分辨毫米波雷达与摄像信息融合策略的目标跟踪算法。首先对毫米波雷达与摄像目标跟踪结果进行信息时空同步,然后在决策级层面采用自适应阈值匹配方法实现目标匹配,按设定的信息融合策略实现传感器信息优势互补。最后结合反馈机制,采用毫米波雷达和摄像检测跟踪一体化方法提高融合跟踪的鲁棒性。大量实测实验表明,所提算法跟踪准确率高于单传感器方案和经典雷视决策级融合方案,且对目标跟踪连续性、鲁棒性具有明显性能改善。

To improve the radar photoelectric target perception ability in radar perception countermeasures,the target tracking algorithm of high-resolution millimeter-wave radar and camera fusion is studied.Aiming at the performance limitations of a single sensor and the weak robustness of the existing radar photoelectric fusion tracking,a target tracking algorithm based on the fusion strategy of high-resolution millimeter-wave radar and camera information is proposed.Firstly,the target tracking results of millimeter-wave radar and camera undergo information space-time are spatiotemporally synchronized.Subsequently,at the decision-making level,adaptive threshold matching method is used to achieve target matching,and sensor information complementary advantages are achieved according to the set information fusion strategy.Finally,combined with the feedback mechanism,a radar-camera detection and tracking integration method is adopted to improve the robustness of fusion tracking.Real measured experiments demonstrate that the proposed algorithm achieves higher tracking accuracy than the single-sensor scheme and classic radar-camera fusion scheme with better performance on the continuity and robustness of target tracking.