基于两类神经网络算法在输电线路中的缺陷识别研究  

Research on Defect Recognition of transmission lines based on two kinds of neural network algorithms

作　　者：张胜 ZHANG Sheng(College of Geomatics and Geographic Information,Lanzhou Jiaotong University,Lanzhou 730070,China;National and Local Joint Engineering Research Center of Geographical Monitoring Technology Application,Lanzhou 730070,China;Gansu Provincial Engineering Laboratory of Geographical Monitoring,Lanzhou 730070,China)

机构地区：[1]兰州交通大学测绘与地理信息学院,甘肃兰州730070 [2]地理国情监测技术应用国家地方联合工程研究中心,甘肃兰州730070 [3]甘肃省地理国情监测工程实验室,甘肃兰州730070

出　　处：《长江信息通信》2023年第9期201-204,共4页Changjiang Information & Communications

基　　金：国家自然科学基金(42101446);中国博士后科学基金(2022T150488);兰州交通大学基础研究拔尖人才培养计划项目(2022JC39)。

摘　　要：针对无人机输电线路巡检图像判读这一问题,文章采用典型的Twostage Cascade R-CNN和Onestage CenterNet两种算法模型分别对绝缘子爆片、鸟窝、线上异物、线上悬挂气球、线上悬挂风筝5类输电线路缺陷进行了识别实验。结果表明,两种模型对选择的5类缺陷的识别平均值均能维持一个较高水平,且CenterNet算法模型的各项指标更好,在后续实际应用中优势更加突出。Aiming at the problem of image interpretation of UAV transmission line inspection,this paper uses two typical algorithm models Twostage Cascade R-CNN and Onestage CenterNet to identify 5 types of transmission line defects,such as insulator explosion,bird nest,foreign matter on line,balloon hanging on line and kite hanging on line respectively.The results show that the two models can maintain a high level of average recognition of the 5 types of defects selected,and the CenterNet algorithm model has better indicators,and has more prominent advantages in subsequent practical applications.

关 键 词：输电线路 神经网络算法 缺陷识别 Cascade R-CNN CenterNet 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]