基于高光谱技术的陶瓷绝缘子污秽等级检测  

Contamination grades detection method of ceramic insulatorbased on hyperspectral technology

作　　者：张血琴 周志鹏 李谦慧 郭裕钧 吴广宁[1] ZHANG Xueqin;ZHOU Zhipeng;LI Qianhui;GUO Yujun;WU Guangning(School of Electrical Engineering,Southwest Jiaotong University,Chengdu 611576,China)

机构地区：[1]西南交通大学电气工程学院,四川成都611756

出　　处：《电力工程技术》2022年第4期150-155,共6页Electric Power Engineering Technology

基　　金：国家自然科学基金资助项目(51907168);四川省杰出青年科技人才项目(2020JDJQ0039);国家电网有限公司总部科技项目(521104190007)。

摘　　要：陶瓷绝缘子污秽等级检测主要采用等值盐密法、表面污层电导法、泄漏电流法等传统方法,检测过程耗时长、效率低。高光谱技术能够非接触地获取目标图谱信息且信息量丰富,在绝缘子污秽检测方面有较大应用潜力,因此文中提出基于高光谱技术的陶瓷绝缘子污秽等级检测方法。首先,由于陶瓷绝缘子的材质原因,采集其高光谱图像时存在反光现象,因此采用直方图均衡化处理高光谱图像,去除反光干扰;然后,对高光谱图像进行预处理,去除噪声干扰;接着,采用连续投影算法(SPA)对样本谱线进行特征提取,去除冗余信息;最后,根据特征谱线建立支持向量机(SVM)分类模型,实现陶瓷样品的污秽等级划分,其准确率可达95%。Traditional indicators such as equivalent salt deposit density,surface pollution layer conductance and leakage current are mainly used to measure the contamination grades of ceramic insulators,and the detection process is time-consuming and inefficient.Hyperspectral technology can obtain rich information of target spectrum non-contact,and has great application potential in insulator pollution detection.Therefore,a method of ceramic insulator pollution level detection based on hyperspectral technology is proposed in this paper.Firstly,there is a reflection phenomenon when collecting hyperspectral image due to the material of ceramic insulator,so histogram equalization is used to process the hyperspectral image to remove the reflective interference.Then,the hyperspectral image is preprocessed to remove the noise interference.And the successive projections algorithm(SPA)is used to extract the feature of the sample spectral line for removing the redundant information.Finally,the classification model of support vector machine(SVM)is established according to the characteristic spectral lines,and the classification accuracy of ceramic samples is 95%.

关 键 词：陶瓷绝缘子 高光谱技术 直方图均衡化 连续投影算法(SPA) 污秽等级检测 支持向量机(SVM) 

分 类 号：TM855[电气工程—高电压与绝缘技术]