Structure defect rapid identification and classification method based on lightweight deep learning model

Abstract

A structure defect rapid identification and classification method based on a lightweight deep learning model comprises the following steps: carrying out semantic segmentation processing on a collected defect image by using a VGG16-U-Net model, removing image background noise interference, constructing an EfficentNetB0 model, training the EfficentNetB0 model to obtain a trained EfficentNetB0 model, and carrying out rapid identification and classification on the defect image by using the VGG16-U-Net model. And inputting the defect image subjected to semantic segmentation processing into the trained OfficientNetB0 model for identification, and outputting an identification result, the method and the device cooperate with a VGG16-U-Net semantic segmentation technology, can realize automatic preprocessing of the image, and improve the robustness of an image identification algorithm. The performance of the data evaluation model is quite accurate, and the model scale and the training speed are well balanced. The training speed and the recognition speed of the model are accelerated, a random gradient descent algorithm of cosine annealing is used to improve the learning rate of the model, the local minimum is avoided, and the global minimum is quickly searched.

Description

technical field [0001] The invention relates to the technical field of defect identification, in particular to a method for quickly identifying and classifying structural defects based on a lightweight deep learning model. Background technique [0002] With the rapid development of the economy, the construction speed of various buildings is getting faster and faster, such as: buildings, bridges, dams and various industrial buildings. At the same time, these buildings will also be damaged during long-term use. Therefore, it is necessary to regularly inspect and repair the building to prevent safety accidents. At present, the most common method is the manual monitoring method based on vision, which uses visual inspection, manual drawing, etc. to identify defects such as cracks in the building. And record its distribution and shape. However, manual inspection has low timeliness and low cost performance. There are also differences in the judgment of the structure, and the accur...

AI Technical Summary

Problems solved by technology

[0004] The embodiment of this application provides a fast identification and classification method for structural defects based on a lightweight deep learning model, aiming to solve the problem of low accuracy of existing image defect identification technologies

Images