Research On Bridge Crack Detection Based On Image Recognition

High speed train is the main tool of today’s passenger dedicated line. As the main carrier of the high-speed railway, the safety of the bridge is very important for the transportation. Bridge crack detection is an important part of bridge detection. At present, the detection of bridge cracks mainly relies on manual testing, which is time-consuming, and the risk is high. Bridge crack detection technology based on image recognition has improved these deficiencies, and become a very active research topic.This paper proposes an improved algorithm for crack identification and crack measurement in the fracture images based on the digital image processing technology. In terms of crack identification, the main work and innovation points are as follows:(1) Through the projection method to carry on the horizontal vertical projection to the crack image, analyzes the characteristic of the projection image, and preliminarily determines the target area containing the crack information in the image; (2) The regional growth of the target area is initially determined, the growth of the regional image contains only the crack information and the background information, so as to achieve the crack identification; (3) The projection method and region growing combined, improve the problem of traditional fracture image recognition process is more cumbersome, directly from the crack area of image segmentation and recognition speed, high degree of accuracy for practical crack detection.In terms of crack measurement, the main work and innovation points are as follows:(1) Through the projection method horizontal, vertical and diagonal cracks are projected horizontally, the number of pixels in the image is statistically projected, and the crack width is calculated by the actual size of each pixel; (2) Projection of the mesh cracks, and calculate the area of the mesh and the space occupation rate; (3) According to the algorithm of the actual crack image classification measurement and statistical data, compared with the traditional manual measurement, analysis of the advantages and disadvantages of the algorithm.This paper improve the algorithm of crack identification and crack measurement based on image processing technology of crack detection. Through the analysis of the experimental results, the method has certain advantages in speed and accuracy, and it is more effective to crack detection of the bottom surface of the bridge.