摘要
绝缘子自爆缺陷检测对于保障输电线路的安全具有十分重要的作用,准确快速检测算法能够帮助运维人员快速定位自爆缺陷绝缘子的位置,并及时更换。传统的人工检测方法已无法满足检测的要求,面向图像的绝缘子自爆缺陷检测算法在其检测的准确性和快速性上仍面临着极大挑战,必须进一步对算法进行改进。本文首先介绍了绝缘子自爆缺陷图像的预处理过程,包括了图像分割的方法以及其具体的算法;其次,介绍了绝缘子自爆缺陷图像的特征提取算法、当前绝缘子自爆缺陷检测常用的分类器以及深度学习网络模型;最后,对绝缘子自爆缺陷检测算法进行了总结,并对其发展进行了展望。
Insulator self-explosion defect detection plays a very important role in ensuring the safety of transmission lines.Accurate and fast insulator self-explosion defect detection algorithm can help operation and maintenance personnel quickly locate the location of self-explosion defective insulators and replace them in time.The image-oriented insulator self-explosion defect detection algorithm still faces great challenges in its detection accuracy and rapidity.The traditional manual detection method cannot meet the detection requirements,so the algorithm must be further improved.The main work of this paper is as follows:firstly,the preprocessing process of insulator self-explosion defect image is introduced,including the method of image segmentation and its specific algorithm;Secondly,the feature extraction algorithm of insulator self-explosion defect image is introduced;Then,the classifiers commonly used in insulator self-explosion defect detection and the deep learning network model are introduced;Finally,the insulator self-explosion defect detection algorithm is summarized,and the development of insulator self-explosion defect detection algorithm is prospected.
作者
赵庆林
陈湘萍
ZHAO Qinglin;CHEN Xiangping(College of Electrical Engineering,Guizhou University,Guiyang,Guizhou 550025,China)
出处
《智能计算机与应用》
2022年第3期33-39,共7页
Intelligent Computer and Applications
基金
国家自然科学基金(51867007)
关键词
深度学习
绝缘子自爆缺陷
特征提取
分类
deep learning
machine vision
feature extraction
classification
