适用于无人机的自动跟踪算法的研究

Research on Automatic Tracking Algorithms for UAVs

入侵农田问题给试验田的安全保护带来了严重挑战,传统的保护手段存在诸多限制。为解决这一问题,将无人机技术和目标检测与跟踪算法相结合,提出一种创新的解决方案。该方法通过无人机高空航拍视角获取农田图像数据,并利用YOLO(You Only Look Once)算法实现实时目标检测。同时,采用SORT(Simple Online and Realtime Tracking)算法对入侵目标进行持续跟踪。通过在海南试验田中的应用实验验证该方法的可行性和有效性。实验结果表明,基于YOLO和SORT算法的无人机目标检测与跟踪系统能够在0.4 s内快速检测和跟踪入侵农田目标,为试验田的安全保护工作提供了重要支持。

Farmland intrusion poses a serious challenge to the security protection of experimental fields,and traditional protection means have many limitations.To solve this problem,an innovative solution is proposed by combining UAV(Unmanned Aerial Vehicle)technology with target detection and tracking algorithms.The method acquires farmland image data through an overhead aerial view of the UAV and realizes real-time target detection using the YOLO algorithm.Meanwhile,the SORT(Simple Online and Realtime Tracking)algorithm is used to continuously track the intrusion targets.The feasibility and effectiveness of the method are verified through application experiments in Hainan test fields.The experimental results show that the UAV target detection and tracking system based on YOLO and SORT algorithms can rapidly detect and track intruding farmland targets within 0.4 s,which provides important support for the protection of the experimental fields.