Model training method, railway overhead line system anomaly detection method and related device

Abstract

The invention discloses a model training method, a railway overhead line system anomaly detection method and device, electronic equipment and a computer readable storage medium. The model training method comprises the steps that an initial training image and a corresponding initial training label are acquired; identifying and cutting a label image corresponding to the initial training label in the initial training image; performing data enhancement on the plurality of categories of label images to obtain enhanced images; wherein the sum of the number of the label images and the number of the enhanced images corresponding to each category is the same; for each initial training image, obtaining a plurality of corresponding target label images and / or target enhanced images, and superposing the initial training images by using the target label images and / or the target enhanced images to obtain corresponding training images; generating a training label corresponding to the training image, and training the initial model by using the training image and the training label to obtain an image processing model; the image processing model obtained by the method has high recognition capability and accuracy.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular to a model training method, a railway catenary abnormality detection method and related devices. Background technique [0002] The high-speed railway catenary is used to provide power for trains, and its operating status directly affects the normal operation of the high-speed railway. In order to ensure the normal operation of the catenary, the relevant railway departments have adopted various methods to inspect the catenary and carry out maintenance work in a timely manner. Currently, maintenance vehicles and other equipment are usually used to collect non-contact images of the catenary, and then use the network model to detect abnormalities in the collected images. However, the size of the abnormal part in the railway catenary image is small, accounting for a small proportion of the entire image, which makes the model unable to effectively learn the characteris...

AI Technical Summary

Problems solved by technology

However, the size of the abnormal part in the railway catenary image is small, accounting for a small proportion of the entire image, which makes the model unable to effectively learn the characteristics of the abnormal part; the probability of anomalies in the catenary is low, which makes the number of training images small; various There are differences in the amount of data between categories of anomalies, that is, the number of anomalies in each category varies greatly

This makes the recognition accuracy of the network model poor

Images