【摘要】 检测并及时修复输电线路的缺陷是电能安全输送的重要保障。针对现有检测方法存在效率低、对多尺度目标检测精度低、泛化能力差等不足，提出了一种基于无人机影像的无锚框输电线缺陷检测方法。该方法基于YOLO系列目标检测框架构建了一种无锚框的检测网络，设计了相匹配的正负样本分配方式，融入了多种优化策略，有效改善了现有方法的不足。实验结果表明，提出的方法能够同时对输电线的断股、散股、断线、烧伤和异物5种缺陷进行有效检测。相比于传统输电线缺陷识别方法和基于深度学习的缺陷检测方法SSD、Faster R-CNN、YOLOv4、YOLOv5,该方法的平均精度均值(mAP)达到78.31%,每秒传输帧数(FPS)为103.5 f/s,同时兼备检测的快速性和高精度，在5类输电线缺陷检测任务中均具有良好的性能。更多还原

【Abstract】 The detection and timely repair of defects in transmission lines is an important guarantee for the safe transmission of electric energy.In view of the shortcomings of the existing detection methods, such as low efficiency, low detection accuracy for multi-scale targets, and poor generalization ability, a kind of unmanned aerial vehicle(UAV) based image without anchor box transmission line defect detection method is proposed.The method is based on YOLO series target detection framework to build a kind of anchor box detection network, the match way of distribution of positive and negative samples is designed, a variety of optimization strategies are integrated to effectively improve the shortcomings of the existing methods.The experimental results show that the proposed method can detect the five defects of broken strand, stray strand, broken wire, burn and foreign body effectively at the same time.Compared with traditional transmission line defect identification methods and deep learning-based defect detection methods SSD,Faster R-CNN,YOLOv4 and YOLOv5,the mean average precision(mAP) of the proposed method is 78.31%,and the frames per second(FPS) is 103.5 f/s.At the same time, it has high detection speed and high precision, and has good performance in five kinds of transmission line defect detection tasks.更多还原