无人机巡检绝缘子智能检测算法及实现  

Intelligent detection algorithm and implementation of UAV inspection insulator

作　　者：谢应昭 左迅 刘健 邓雅文 谢秋洪 XIE Ying-zhao;ZUO Xun;LIU Jian;DENG Ya-wen;XIE Qiu-hong(The Information and Telecommunication Branch of State Grid Chongqing Electric Power Company,Chongqing 400012,China;School of Electrical and Electronic Engineering,Chongqing University of Technology,Chongqing 400054,China;The Information and Telecommunication Branch of Chongqing Guanghui Electric Power Service Supply Company,Chongqing 400014,China)

机构地区：[1]国网重庆市电力公司信息通信分公司,重庆400012 [2]重庆理工大学电气与电子工程学院,重庆400054 [3]重庆广汇供电服务有限责任公司信息通信分公司,重庆400014

出　　处：《信息技术》2023年第7期64-70,76,共8页Information Technology

基　　金：重庆市基础科学与前沿技术研究专项基金资助项目(cstc2019jcyj-msxmX0233);重庆市教育委员会科学技术研究计划基金资助项目(KJQN201901125)。

摘　　要：针对电网巡检中人工识别绝缘子工作量大、缺陷检测实时性差的问题,提出基于Faster R-CNN的无人机目标快速检测算法,通过将绝缘子的样本进行特征提取,形成不同尺度的检测通道,并对缺陷绝缘子样本集进行匹配,完成绝缘子的快速识别和检测。将所提算法移植于无人机计算平台进行实际验证,结果表明:所提算法可在35ms内完成识别和检测绝缘子,正确率达90%,可满足电网运检智能化、自动化的要求。Based on the problems of heavy workload in insulator artificial identification and poor real-time defect detection,the target fast detection algorithm of UAV based on Faster R-CNN is proposed.It extracts the key features of insulator in order to making different scale detection channels,and accomplishes fast identification and inspection of insulator by matching the defect insulator sample sets.The algorithm has been transplanted to UAV computing platform in order to verificate.The result shows that the proposed algorithm can identificate and inspect insulator in 35ms,reaching the 90%accuracy,which can meet the requirement of the intelligence and automation in power grid inspection.

关 键 词：无人机巡检 绝缘子 目标检测 图像识别 Faster R-CNN 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]