Automatic detection of transformer respirator based on image processing

Abstract

The traditional transformer respirator lacks the detection means, cannot realize the long-distance monitoring, and must carry on the regular inspection through the running staff to determine its state. This paper introduces a method of automatic detection of transformer respirator based on image processing with the aid of the intelligent inspection robot on-site shooting, returning the transformer respirator device image, using the digital image processing technology for color histogram analysis of the picture, and last making decisions to determine whether the need to replace the discoloration of silica gel. This method not only does not require any requirement for the structure of the transformer respirator, but also does not require any sensor device, thus greatly reducing the detection cost of the transformer respirator, both convenient and practical.