基于机器视觉的电缆局部绝缘材料老化缺陷检测研究  

Research on Aging Defect Detection of Cable Local Insulation Materials Based on Machine Vision

作　　者：吴伟 查姿伊 刘晔 朱鉴恺 WU Wei;ZHA Zi-yi;LIU Ye;ZHU Jian-kai(Changzhou Power Supply Company of State Grid Jiangsu Electric Power Co.,Ltd.,Changzhou 213000 China)

机构地区：[1]国网江苏省电力有限公司常州供电分公司,江苏常州213000

出　　处：《自动化技术与应用》2024年第6期156-160,共5页Techniques of Automation and Applications

基　　金：国网江苏电力有限公司科技项目(FZJS2200479)。

摘　　要：研究基于机器视觉的电缆局部绝缘材料老化缺陷检测方法,以期实现电缆局部绝缘材料老化缺陷准确检测。使用基于机器视觉技术的材料图像采集方法,通过CCD相机与光学镜头相互配合采集材料图像,并使用畸变校正模型,控制图像像素不出现过分畸变状态;由基于逆直方图均衡化的绝缘材料图像增强算法,提高电缆局部绝缘材料图像清晰度后,经基于改进深度卷积神经网络的绝缘材料老化缺陷识别方法,提取增强后绝缘材料图像的目标特征,识别目标特征是否属于老化缺陷类型,完成电缆局部绝缘材料老化缺陷检测。研究结果显示方法对电缆局部绝缘材料老化缺陷检测结果准确,且具备较好的图像处理效果。The aging defect detection method of cable local insulation materials based on machine vision is studied in order to realize the accurate detection of aging defects of cable local insulation materials.Using the image acquisition method of cable local insulation material based on machine vision technology,the image of cable local insulation material is collected through the cooperation of CCD camera and optical lens,and the distortion correction model is used to control the image pixels from excessive distortion.After the insulation image enhancement algorithm based on inverse histogram equalization improves the definition of the local insulation image of the cable,through the insulation aging defect identification method based on the improved depth convolution neural network,the target features of the enhanced insulation image are extracted,whether the target features belong to the aging defect type is identified,and the aging defect detection of the local insulation of the cable is completed.The results show that the method is accurate in detecting the aging defects of cable local insulation materials,and has a good image processing effect.

关 键 词：机器视觉 电缆 绝缘材料 老化缺陷检测 目标检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP391.413[自动化与计算机技术—计算机科学与技术]