基于PK-ByteTrack目标跟踪算法在无人机探测与跟踪中的应用OA
Application of PK-ByteTrack object tracking algorithm in UAV detection and tracking
为提高无人机在复杂红外场景下的跟踪精度与稳定性.本文提出了一种改进的目标跟踪算法 PK-ByteTrack,该算法采用精度更高的 PCCS-YOLOv8 检测器,并对卡尔曼滤波器的状态向量与噪声模型积极性优化.经研究表明,PK-ByteTrack 算法的 MOTA 平均提升3.8%,IDSW 平均降低22.6%,同时在 FP 与 FN 控制方面表现均优于其他算法.该算法在保持实时处理能力的同时,显著提升在复杂场景下对无人机目标的跟踪鲁棒性与轨迹连续性.
To improve the tracking accuracy and stability of unmanned aerial vehicles(UAVs)in complex infrared sce-narios,this paper proposes an improved target tracking algorithm named PK-ByteTrack.The algorithm employs the more accurate PCCS-YOLOv8 detector and optimizes the state vector and noise model of the Kalman filter.Research re-sults show that the PK-ByteTrack algorithm improves the MOTA by an average of 3.8%and reduces the IDSW by an average of 22.6%,with superior performance in FP and FN control compared with other algorithms.While maintaining real-time processing capability,the algorithm significantly enhances the tracking robustness and trajectory continuity for UAV targets in complex scenarios.
李沙
山西通用航空职业技术学院,山西 大同 037000
航空航天
无人机目标跟踪ByteTrack卡尔曼滤波多目标跟踪
UAV target trackingByteTrackKalman filteringmulti-object tracking
《农机使用与维修》 2026 (6)
7-10,4
评论