Damage detection method based on porcelain insulator

Abstract

The invention provides a damage detection method based on a porcelain insulator, and the method comprises the steps: step 1, selecting 1500 images photographed by an unmanned plane as original images,positioning the porcelain insulator in the original images through YoLoV3, segmenting the positioned porcelain insulator from the original images, and obtaining a porcelain insulator image; step 2, extracting the contour of the porcelain insulator in the porcelain insulator image by using DeepLabV3; and step 3, carrying out median filtering and binarization preprocessing on the extracted contourof the porcelain insulator, finding out the maximum contour in the same porcelain insulator image, and calculating the area and hu moment of the maximum contour. According to the invention, accurate damage detection can be carried out on porcelain insulator images shot at different angles; the damage condition of the porcelain insulator can be judged more accurately through multi-angle shooting, the requirement for unmanned aerial vehicle shooting is not high, the damage detection accuracy of the porcelain insulator is improved, errors caused by DeepLabV3 contour extraction are effectively avoided, and the method has very high application value.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a damage detection method based on an insulating porcelain bottle. Background technique [0002] At present, target detection based on deep learning is becoming more and more popular, and its accuracy is very high. More and more people use YoLoV3 to achieve target detection. YoLoV3 is based on the darknet framework, which uses pure C language and does not rely on other third-party libraries. The caffe framework is friendly to developers in terms of ease of use. [0003] The DeepLabV3 series proposed by the Google team is a model that is convenient and widely used for image semantic segmentation, and its latest DeepLabV3 architecture mainly uses DeepLabV3 as the encoder architecture. The decoder uses a simple but effective module, and explores improved Xception and depth separation. The application of convolution in the model further improves the performance of the mode...

AI Technical Summary

Problems solved by technology

[0005] The invention provides a damage detection method based on an insulating porcelain bottle, the purpose of which is to solve the problem that the traditional detection method of an insulating porcelain bottle requires a high shooting angle of the insulating porcelain bottle and the detection accuracy is not high

Images