摘要
为了提高无人机自主精准降落的准确性、增强无人机自主着陆的适应性,研究了一种新式无人机跟踪算法。首先,对无人机与着陆目标确定相对位置的计算过程和原理进行分析,总结出传统着陆目标跟踪算法的缺点。然后,创新性地将跟踪学习检测(TLD)算法与目标跟踪中的核化相关滤波(KCF)算法相结合,利用KCF算法的优势优化TLD算法,得到TLD+KCF目标跟踪算法。最后,提出基于无人机降落的优化算法,并设置对照组验证算法性能。对比结果表明,所提算法的准确率和成功率超过了对比算法。该算法精度高、稳定性强,可实现无人机自主精准降落。该研究有助于提高无人机自主精准降落的准确性。
To improve the accuracy of autonomous and precise landing of unmanned aerial vehicles and enhance the adaptability of autonomous landing of unmanned aerial vehicles,a new type of unmanned aerial vehicles tracking algorithm is studied.Firstly,the computational process and principle of determining the relative position between unmanned aerial vehicles and landing target are analyzed,and the shortcomings of the traditional landing target tracking algorithm are summarized.Then,the tracking learning detection(TLD)algorithm is innovatively combined with the kernelized correlation filtering(KCF)algorithm in target tracking,and the advantages of the KCF algorithm are used to optimize the TLD algorithm to obtain the TLD+KCF target tracking algorithm.Finally,the optimization algorithm based on unmanned aerial vehicles landing is proposed,and a compared group is set to verify the performance of the algorithm.The comparison results show that the accuracy and success rate of the proposed algorithm exceeds that of the comparison algorithm.The algorithm is highly accurate and strongly stable and can realize autonomous and precise landing of unmanned aerial vehicles.The study helps to improve the accuracy of autonomous and precise landing of unmanned aerial vehicles.
作者
陈潇
徐曙
钟灿堂
赵晓丹
CHEN Xiao;XU Shu;ZHONG Cantang;ZHAO Xiaodan(Shenzhen Power Supply Bureau Co.,Ltd.,Shenzhen 518000,China;Guangzhou Zhongke Yuntu Intelligent Technology Co.,Ltd.,Guangzhou 510095,China)
出处
《自动化仪表》
CAS
2024年第9期65-69,75,共6页
Process Automation Instrumentation
关键词
无人机
自主降落
控制系统
跟踪学习检测算法
核化相关滤波算法
目标追踪
Unmanned aerial vehicles
Autonomous landing
Control system
Tracking learning detection(TLD)algorithm
Kernelized correlation filtering(KCF)algorithm
Target tracking
