Building surface crack detection method based on deep learning network

Abstract

The invention discloses a building surface crack detection method based on a deep learning network. The method comprises the following steps: creating a network training data set; constructing a deeplearning network model, and performing training; detecting an image by using the trained deep learning network model, and outputting a prediction label image; carrying out analysis and detection basedon crack characteristics on the prediction label image containing the crack; and calculating and outputting crack parameters, including coordinate information of each crack, and length and width of each crack. Compared with a traditional crack detection method, the method has the advantages that the thought of the full convolutional network and the thought of the residual network are combined, the deep learning network used for crack detection is constructed, color images of any size can be used as input, end-to-end detection is achieved, and the model has high detection accuracy and high generalization ability; besides, the cracks are connected based on the characteristics, the length and the width of the cracks are calculated, and a powerful means is provided for quantitative evaluationof the cracks.

Description

technical field [0001] The invention belongs to the field of engineering defect detection, and in particular relates to a method for detecting cracks on a building surface based on a deep learning network. Background technique [0002] The regular inspection and maintenance of roads, bridges, dams, high-rise buildings and other buildings is an important prerequisite and guarantee to ensure their safe operation. Crack defect detection is one of the important items. However, at present, manual inspection has problems such as low efficiency, incomplete data statistics, and high difficulty in ensuring the safety of inspectors. Therefore, automated methods are urgently needed to realize inspection of crack defects. UAVs, robots and other equipment carry image acquisition devices, which provide advanced means for automatic data collection, so the automatic detection algorithm of the collected images becomes the bottleneck of the system, which is crucial to the automation and accu...

AI Technical Summary

Problems solved by technology

[0008] Aiming at the problem that the current crack detection network model is not superior enough, and it is difficult to obtain the shape, length and width of cracks, the present invention provides a crack detection method on the surface of buildings based on a deep learning network to improve the accuracy and efficiency of crack detection

Images