Contamination condition evaluation method based on insulator image feature dictionary

Abstract

The invention relates to a contamination condition evaluation method based on an insulator image feature dictionary, which comprises three parts of insulator image feature dictionary construction, region segmentation and classifier training. The feature dictionary is constructed by utilizing sample insulator images; for a shot insulator under a complex background, by comparing to the feature dictionary, an identified insulator region is extracted; and by classifier training, discrimination on different contamination degrees of insulators is implemented. Compared to the prior art, the method disclosed by the invention has the advantages that identification accuracy is high, and the method is convenient for a worker to operate, is beneficial for reducing an insulator contamination flashoverpower failure and can be used for insulator contamination condition detection places and the like under the complex background of transformer substation and power transmission and distribution line polling and the like.

Description

technical field [0001] The invention relates to the fields of electrical engineering and computer vision, in particular to a pollution state assessment method based on an insulator image feature dictionary. Background technique [0002] With the deterioration of the atmospheric environment, the pollution of insulators is serious, and the electric strength is greatly reduced, which is easy to cause pollution flashover accidents and seriously threatens the safety of the power system. An effective measure to prevent pollution flashover of insulators is to accurately judge the pollution level of insulators. There are many kinds of insulator pollution level discrimination methods, mainly including voltage distribution detection method, leakage current method, pulse current method, spectroscopy method, ultraviolet imaging method, infrared thermal imaging method, visible light image recognition method, etc. Among them, image-based insulator pollution identification methods such as...

AI Technical Summary

Problems solved by technology

Common visible light image recognition directly divides insulators into regions. Due to the high similarity of different types and different types of insulators, it is easy to cause misidentification, and usually the background of the picture where the insulator is located is complex and there are many interference objects. Common recognition methods are easy to cause the recognition frame to be divided Inaccurate

Images