An Adaptive Diagnosis Method for Porcelain Insulator Deterioration Based on Infrared Images
To address the challenge of diagnosing infrared image degradation of porcelain insulators,this paper proposes an adap-tive intelligent diagnosis method.First,high-precision location of the entire insulator string and steel caps is achieved through two deep learning networks of varying complexity.After calculating the relative temperature-difference curve for the entire insu-lator string according to industry standard,the Local Outlier Factor(LOF)algorithm is employed to adaptively determine the defect threshold.This method is designed to be applicable to infrared images captured by different models of thermal imagers.The experimental section analyzes key parameters and demonstrates the effectiveness of the method,achieving a degradation di-agnosis accuracy of over 90％.This approach significantly enhances the efficiency and effectiveness of porcelain insulator degra-dation diagnosis.
万方数据
维普
