Damage detection method of carbon fiber composite core cable based on image classification network

Abstract

The invention provides a damage detection method for carbon fiber composite core cables based on an image classification network, which uses X-imaging to collect images of carbon fiber composite core cables; performs preprocessing operations on the collected images of carbon fiber composite core cables, and obtains training samples and detection methods respectively. Sample; determine the input and output of the initial residual network and the structure of the middle part, and construct the carbon fiber composite core cable damage classification detection network; use the training samples to train the carbon fiber composite core cable damage classification detection network, and save the model with the best training effect; use The detection sample is tested, and the damage position in the image is automatically marked according to the classification output of the carbon fiber composite core cable damage classification detection network. The method can automatically realize the damage detection of the carbon fiber composite core cable, and has high detection efficiency and precision.

Description

technical field [0001] The invention relates to a method for detecting damage of a carbon fiber composite core cable based on an image classification network. Background technique [0002] With the continuous progress of modern industries, the power consumption of all walks of life is increasing geometrically, and the load on cables is increasing. Due to the complex and diverse environment of the cable, whether it is construction or use, it is easy to cause some damage to the carbon fiber composite core, which directly affects the normal power supply and greatly affects the lives of residents. Therefore, the detection method of the cable damage position is particularly important. [0003] In the traditional flaw detection method, Wan Jiancheng of China Electric Power Research Institute and Zhu Bo of Shandong University and others mainly used three kinds of detection techniques: ray detection, frequency analysis and sag monitoring. After communicating with units that have a...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a carbon fiber composite core cable damage detection method based on an image classification network to solve the problem of carbon fiber composite core cable damage detection that is difficult to detect due to the small difference between the background and damage in the prior art

Images