The Research On Rapid Detection Of Characteristics Of Tunnel Lining Surface Diseases

With the current construction of a large number of high-speed railway owned high tunnel-line ratio, The status that a large number of tunnels need to be tested and maintained will come in the future. However traditional detection methods with a low efficiency can not meet the detecting demands of high-speed railway. Therefore, it is necessary to study a technology programs for rapid detection of tunnel lining surface crack diseases. The detection method based on image processing is used to study the extraction method of surface crack image in the tunnel lining, mainly involving five main aspects about the detection method, which includ image correction, image denoising, image segmentation, cracks quickly identify and disease detection.For the distortion of image acquisition, the image is adjusted based on the principle of distortion in this thesis, so that the image information is more accurate. When image is shooting, researchs on the change of the pixel equivalent are done and corresponding correction method and suggestion are proposed.In image processing, Adaptive value-Wiener filtering method is used for processing image denoising, which preserves much fringe information of the image. Because the gray values of tunnel image background are in chaos, an Otsu method which is improved by image edge imformation is used to segment image; and impurities area are reduced in the image by outlier removal and morphological processing to the segmented image.In the aspect of crack quikly identify, because of the large number of images collected by the detection system, and the divided image area still containing impurities, cracks can not be extracted directly. Based on the difference between crack and non-crack feature, six features which include length area, occupancy, average width, the number of background connected domain and roundness,are selected to establish a recognition model by using the support vector machine method, enabling rapid identification of cracks in the image area.In the detection of disease detection in tunnels, the main item are the detection of development of cracks and block area. Through the skeleton image stitching and Mileage positioning, the contrast of crack image taken at different times comes true, which contribute to researching the development of crack. Through analyzing the cracks skeleton in complex crack regions, the method of loop detection is proposed to determine whether the crack region contains block area.This thesis is mainly based on image processing of surface disease detection in the tunnel, trying to establish a complete process for collection of image processing and detection of the disease, establishing the foundation of the tunnel structure surface disease detection systems.