基于图像识别技术的变电站二次保护状态检测方法  

Detection Method of Substation Secondary Protection Status Based on Image Recognition Technology

作　　者：常荣 杨传旭 CHANG Rong;YANG Chuan-xu(Yuxi Power Supply Bureau of Yunnan Power Crid Co.,Ltd.,Yuxi 653100 China)

机构地区：[1]云南电网有限责任公司玉溪供电局,云南玉溪653100

出　　处：《自动化技术与应用》2024年第12期175-177,182,共4页Techniques of Automation and Applications

基　　金：国家自然科学面上项目(41971392)。

摘　　要：变电站采用人工检测判断设备的方式存在作业效率低,工作质量不高等问题,且受光照变化的影响,导致检测结果会出现偏差。为此,设计基于图像识别技术的变电站二次保护状态自动化检测方法。在图像采集过程中,在待检测区域将工控摄像头均匀布置,需在不同时间点对待检测位置进行拍摄,实现同步拍摄,利用中值滤波和AND原理去除图像噪声,设置特征提取窗口,利用二值化处理设置灰度值比较函数,建立特征编码模型,提取Haar特征,找到图像中匹配的状态模型并进行几何形态学操作,判断当前的二次保护设备状态是否正常。测试结果表明:不同光照条件下,所提方法的检测结果轮廓更加清晰,检出率和误检率的结果优于传统方法。The substation adopts manual detection to judge the equipment,which has the problems of low operation efficiency and low work quality,and is affected by the change of light,resulting in the deviation of detection results.Therefore,an automatic detection method of substation secondary protection status based on image recognition technology is designed.In the process of image acquisition,the industrial control cameras are evenly arranged in the area to be detected,and the positions to be detected at different time points need to be photographed to achieve synchronous shooting.The median filter and the and principle are used to remove the image noise,the feature extraction window is set,the gray value comparison function is set by the binarization process,the feature coding model is established,the Haar feature is extracted,the matching state model in the image is found,and the geometric morphology operation is performed,Judge whether the current secondary protection equipment status is normal.The test results show that under different illumination conditions,the detection results of this method are more clear,and the detection rate and false detection rate are better than the traditional method.

关 键 词：图像识别 变电站 二次保护 状态检测 HAAR特征 滤波去噪 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] TP391.413[自动化与计算机技术—控制科学与工程]