Concrete crack identification method based on YOLOv3 deep learning

Abstract

The invention belongs to the technical field of concrete structure damage detection, and discloses a multi-target crack recognition method based on a YOLOv3 deep learning algorithm, which comprises the following steps: importing a crack image into a YOLOv3 model, and automatically compressing the image into 416 * 416 pixel resolution; dividing the original image into S * S grids according to the scale size of the feature map by adopting an up-sampling and feature fusion mode similar to FPN; taking the cross-to-parallel ratio of the candidate box and the real box as an evaluation criterion, and; performing K-means clustering analysis on mark boxes for all crack target marking boxes of the image training set to obtain the size of a candidate box; and predicting the probability that the frame contains the target for each boundary frame through logistic regression. According to the method, the complexity of network training is simplified, and the operation cost is reduced; according to the method, the multiple targets are quickly and accurately identified, the accuracy far superior to that of other models is obtained while the target detection is quickly realized, and the method has higher robustness and generalization capability and is more suitable for an engineering application environment.

Description

technical field [0001] The invention belongs to the technical field of damage detection of concrete structures, in particular to a concrete crack recognition method based on YOLOv3 deep learning. Background technique [0002] Concrete is the most widely used and used building material today, and is widely used in the construction of infrastructure such as roads, bridges, housing, tunnels and dams. Due to the low tensile strength of concrete, affected by both internal and external factors such as shrinkage and creep, external temperature changes, and foundation deformation, cracks of various degrees and forms often occur during construction and operation. The expansion of cracks is the initial stage of structural damage; as the cracks continue to develop, once the width of the crack exceeds a certain limit, it will not only affect the appearance of the infrastructure, but also may cause leakage, reduced durability, peeling off of the protective layer, corrosion of steel bars,...

AI Technical Summary

Problems solved by technology

[0011] To sum up, the problems existing in the existing technology are: the currently commonly used edge detection method is not suitable for the field of crack identification; the deep learning method based on the sliding window technology requires a large amount of data, lacks pertinence, high computing cost, and the detection low efficiency

[0012] Difficulty in solving the above technical problems: There is a lack of recognized and effective crack data sets worldwide, and manual production of data sets requires a lot of manpower and material resources and high costs; simplify the complex model network structure and reduce the intermediate output of the network. To maintain the accuracy and speed of detection, it takes a lot of time to adjust the network, trial calculation, training, etc.

Images