Defect detection method for composite insulator based on infrared image segmentation and SSA-SVM algorithm
DONG Yifei1, SHU Shengwen1, CHEN Cheng1, JIN Ming2, WANG Jian2
1. School of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350108; 2. Electric Power Research Institute, State Grid Xinjiang Electric Power Co., Ltd, Urumqi 830013
Abstract：In order to identify the defect types of AC/DC composite insulator accurately and quickly, a defect detection method for AC/DC composite insulators based on infrared image segmentation and support vector machine (SVM) optimized by sparrow search algorithm (SSA) is proposed. Firstly, short samples of composite insulators are made, and four different types of simulated defects are set. By applying AC and DC voltage respectively, the infrared image samples of normal and defective composite insulators are measured by infrared thermal imager. Then, the insulator region is obtained by a threshold segmentation method based on the maximum variance between classes, and the infrared feature parameters is calculated and selected by the Fisher criterion. Finally, the SVM model optimized by SSA is used to identify the insulator defect types. The results show that the method has a recognition accuracy of more than 87% for the defect types of AC/DC composite insulators in the laboratory and good recognition effect for AC composite insulators in primary field application.
2021, Vol. 22 
