以机器视觉为支撑的玻璃绝缘子缺陷检测技术
Machine Vision-based Defect Detection for Glass Insulators
李乔森 1乔咪咪1
作者信息
- 1. 兰州陇能电力科技有限公司,甘肃 兰州 730000
- 折叠
摘要
为解决电力设备安全中的核心问题即检测绝缘子的非法使用,以机器视觉为支撑,图像获取单元利用 LED红光源垂直照射绝缘子,CCD相机捕获反射光线,通过图像采集卡实时采集信号,而图像处理与缺陷识别模块则使用图像处理软件 MATLAB等,经过相应的图像细化预处理、阈值分割和特征提取,有效检测和分类缺陷.在图像预处理方面,该技术采用了灰度值变换和图像滤波技术,特别改进了高斯低通滤波,以平衡噪声去除和细节保护的需求.随后,通过阈值分割将图像划分为不同区域,并提取关键特征参数,有助于准确辨别和分类不同缺陷.经实验证明,该技术可成功检测气泡和裂缝缺陷.
Abstract
This study aims to address the core issue of power equipment safety,which is the detection of violation in insu-lators usage,and proposes a technology based on machine vision.The image acquisition unit vertically irradiates the insu-lator with an LED red light source and detects the reflected light with a CCD camera,and collects real-time signals through an image acquisition card.The image processing and defect recognition module uses image processing software such as MATLAB,and undergoes corresponding image refinement preprocessing,threshold segmentation,and feature extraction to effectively detect and classify defects.In terms of image preprocessing,the technology adopts grayscale transformation and image filtering techniques,particularly improving Gaussian low-pass filtering to balance the needs of noise removal and detail protection.Subsequently,the image is divided into different regions through threshold segmenta-tion and key feature parameters are extracted,which helps to accurately identify and classify different defects.The tech-nology is verified through experiments capable of achieving detection of bubble and crack defects.
关键词
机器视觉/玻璃绝缘子/缺陷检测Key words
machine vision/glass insulator/defect detection引用本文复制引用
出版年
2024
NETL
NSTL
万方数据