Mask-RCNN-based power equipment infrared image segmentation method

Abstract

The invention discloses a Mask-RCNN-based electric power equipment infrared image segmentation method, and the method comprises the following steps: S1, building a data set of an electric power equipment infrared image, and marking a training set and a test set; S2, constructing a vertical deep learning model; S3, setting initial hyper-parameters and the number of iterations of the model; S4, using the training set marked in the step S1, and inputting the training set into the constructed model for training; s5, evaluating the performance of the model obtained by the training in the step S4 byadopting the test set marked in the step S1 every 2000-3000 iterations; s6, when the number of iterations reaches a set value, stopping training, and screening out a deep learning model with the optimal performance; and S7, inputting the infrared image of the to-be-tested power equipment into the trained optimal deep learning model for processing, and obtaining a segmentation result. According tothe method, the segmentation precision is remarkably improved, the original color information of the target equipment is reserved, the temperature information can be obtained, and a basis is providedfor fault diagnosis.

Description

technical field [0001] The invention relates to an infrared image segmentation method of power equipment based on Mask-RCNN, which belongs to the field of image processing of power equipment in substations. Background technique [0002] In recent years, many monitoring technologies have been actively promoted and applied. Among many monitoring technologies, infrared thermal imaging technology is favored because of its advantages such as no power failure, no sampling, and no disassembly. At present, the analysis and diagnosis of infrared images mostly rely on manual work, and the current infrared monitoring personnel are few, unable to meet the analysis work of huge infrared images. In addition, the professional knowledge level and monitoring experience of some monitoring personnel are insufficient. The analysis and judgment ability of the equipment is poor, which greatly restricts the improvement of the intelligent level of equipment condition monitoring. [0003] In order ...

AI Technical Summary

Problems solved by technology

At present, the traditional image segmentation method is mainly suitable for the situation where the image target is prominent and not dense, and the accuracy of the segmentation result is not high.

In addition, the image information of the original infrared image is not preserved in the current segmentation result map of electric equipment based on the traditional method, and the output result is a black and white image. Since the color information in the infrared image corresponds to the temperature information of the equipment, this is an infrared fault of the electric equipment. An important basis for diagnosis, so the traditional method of power equipment segmentation is not conducive to infrared fault diagnosis and analysis

Images