Defects Detection of Glass Insulator Based on Color Image

Abstract

Helicopter patrol inspection system has been applied to diagnose the insulator injuries of overhead transmission lines.In this paper a method is proposed to improve the diagnosis efficiency.Firstly,the statistical information of blocked image,the form of connected domain and the characteristics of edge chain code are utilized to recognize the area where the defective glass insulator locates;then by use of sliding window histogram statistic and histogram matching judgment,the damaged region of glass insulator is recognized.The proposed method is suitable to real-time detection in field environment,and it can diagnose the injury of glass insulator under a certain light variation range and background complexity.Using the proposed method,the injury of glass insulator can be rapidly and accurately detected at lower omission factor while the real-time processing can be also implemented by airborne system.