Study On The Video Detection System Of Crack On The Bottom Of High-speed Rail Bridge

Cracks in the concrete bridge may lead to serious consequences, so the detection of them is very important in the high-speed rail bridge safety detection. Currently, the artificial detection method is widely used in crack detection. However, there are many disadvantages, such as low efficiency, low Security and high labor intensive, and so on. Therefore, we propose a new detection method of crack in high-speed rail bridge, and also develop a video detection system. Thus we improve current crack detection technology.In this paper, we first collect video images of crack on the bottom of bridge by CCD camera and image acquisition card, and then finish the video image processing based on the video image technology and the BP neural network technology. By two steps mentioned above, we achieve the goal of off-line detection of crack on the bottom of high-speed rail bridge, and keep the precision arrive at 0.2mm.Some key problems are solved in the process of research. To exclude a large number of images without cracks, we propose a scanning method based on the 2nd derivative of image gray for image pre-detection. To distinguish the distractions in the crack images, we use image processing methods which include wavelet filter, edge detection, iterative threshold, morphological processing and removing isolated noise to identify crack and extract feature of crack. Considering the diversity and complexity of the crack, we build a BP neural network for crack classification to provide some convenient for the maintenance of the bridge.Finally, we develop a crack detection system of high-speed railway bridge based on Visual C+ + software. The experiment results show that this system is feasibility and reliable.