Crack Detection Of Bridge On Video Image

Digital Image Processing technology, which is Associated with Mathematics, Pattern Recognition, Artificial Intelligence, Computer Science, Information Theory, Biomedical and other disciplines, is a multidisciplinary application of technology.In this process, the image signal is converted to digital format. And then it uses computer to handle.There are two methods to treat it, including spatial domain processing and transform domain processing.Image technology contains rich. Target detection is one of image segmentation, based on target geometry and statistical characteristics. The accuracy and real-time is one of the important capability of the entire system.Now it widely used. Crack detection of bridge based on digital image processing is a representative application. It can identify the crack fast and accurate and provide the details for health maintenance. In this paper, various image processing algorithms are used,combined with image recognition technology,we developed a bridge Crack Detection System which is simple,fast and accurate,suiting far distance, under the VC++ environment. The major work of this paper are as follows:1) It presents the knowledge needed to use in the image preprocessing of crack detection, such as gray, gray level correction, denoising and traditional edge detection algorithms,and given the corresponding experimental results comparison chart at the same time.We propose the advanced median filter method,which can preserve image details while denoising.And then comparison chart is given,2) Optimizing the eight directions of Sobel edge detection and the experimental results are presented. It can effectively reduce the false edges; Segementing crack image using the maximum variance adaptive threshold which makes a effective results and make feature calculation easy.3) In calculating the crack geometry characteristics, we use the method of extracting skeleton on single crack length. We present a new method on calculation the area of the image by using the template logic and the crack image which is more accurate. It provide more valid data for the bridge health monitoring.4) Based on VC++ we develop a complete set of fast crack detection system.It consists of two parts including image acquisition subsystem and data processing subsystem.The crack image processing subsystem is an integrated application system contains all processing algorithms,and it can achieve very good results.