Crack detection method and device based on semi-supervised semantic segmentation

Abstract

The invention discloses a crack detection method and device based on semi-supervised semantic segmentation. The method mainly comprises the following steps: acquiring a to-be-detected crack image; inputting the crack image into the student model and the teacher model for training, and updating the weight of the student model through gradient descent of a loss function obtained through training; updating the weight of the teacher model through the index moving average value of the weight of the student model; and performing accuracy evaluation on the training model. And the student model and the teacher model have the same network structure taking an EfficientNet as an encoder and a UNet as a decoder, so that multi-scale crack feature information can be efficiently extracted, and meanwhile, the loss of image information is reduced. And a semi-supervised learning means is also adopted, so that the annotation workload is reduced. Experiments show that high detection precision can be maintained while the workload of data annotation is reduced.

Description

technical field [0001] The invention relates to the field of image detection, in particular to a surface crack detection method and device based on a semi-supervised semantic segmentation network. Background technique [0002] With the development of our country's economy and society, most of the infrastructure facilities have suffered varying degrees of wear and tear due to overloading. Among them, the cracks on the surface of the facilities are an obvious sign of facility wear and tear. The detection of surface cracks is crucial to ensure the safety and performance of civil infrastructure. [0003] In recent years, the automatic detection method has gradually replaced the traditional manual detection due to its high efficiency and objective detection results. In the automatic detection method, the semantic segmentation algorithm based on deep learning has shown good performance in crack detection. However, in the semantic segmentation algorithm, the currently commonly us...

AI Technical Summary

Problems solved by technology

However, in the semantic segmentation algorithm, the currently commonly used fully supervised segmentation method needs to manually label a large amount of data, which takes a long time

It is not conducive to the promotion and application in the field of image detection

Images