Fast Detection Algorithm of Tunnel Surface Cracks Based on Image Processing

Cracks are one of the important factors that affect the internal safety of tunnels, and may further deform with the passage of time, which finally has a serious impact on the safety of subway operation, so it is necessary to closely monitor specific dangerous cracks. As a distinguishing structural algorithm, the deep convolution network has excellent performance in image classification and recognition, which is suitable for large sample image processing and can be used for the detection and recognition of complex crack images in subway tunnels. In this paper, a fast detection algorithm of tunnel surface cracks is proposed. The algorithm constructs a tunnel surface crack identification model through convolutional neural network, and removes isolated points and morphologically processes the segmented image to reduce impurity areas in the image. Verified by the data collected in practical engineering, this method has a good effect on crack identification. This detection method can quickly obtain the crack data, which is conducive to the objective evaluation of the structural safety of the tunnel and avoids the interference caused by personal factors of technicians.