摘要
基于路径规划算法与机器视觉技术,对无人机光伏巡检路线进行优化,提出一种基于A*算法路径规划与机器视觉实现的无人机巡检方法。通过确定无人机的最佳飞行路线,使其能够高效并精确覆盖光伏电站的各个区域,并利用改进后的YOLOv5机器视觉技术,对光伏电板进行图像识别和分析,实时检测并诊断潜在的故障或损坏。实验结果表明:优化后的路线使得无人机的巡检效率大大提高,节省了时间和资源;改进后的机器视觉技术对光伏热斑尤其小故障有较高的识别率。
The unmaned aerial vehicle photovoltatic(UAVPV)inspection routes are optimized based on path planning and machine vision technology.A UAVPV inspection based on A*algorithm path planning and machine vision is proposed to determine the best flight route for the UAV so that it can efficiently and accurately cover various areas of the photovoltaic power station.The improved YOLOv5 machine vision technology is used to perform image recognition and analysis on photovoltaic panels to detect and diagnose potential faults or damages in real-time.Experimental results show that the optimized route greatly improves the efficiency of UAV inspections and saves time and resources.The improved machine vision technology has a very high recognition rate for photovoltaic hot spots,especially minor faults.
作者
王云冰
付晓刚(指导)
牛源
WANG Yunbing;FU Xiaogang;NIU Yuan(School of Electrical Engineering,Shanghai Dianji University,Shanghai 201306,China)
出处
《上海电机学院学报》
2023年第5期275-280,共6页
Journal of Shanghai Dianji University
关键词
路径规划
深度学习
神经网络
机器视觉
目标检测
path planning
deep learning
neural networks
machine vision
target detection
