Research Of Vehicle-mounted Tunnel Lining Crack Detection System Based On Imaging Processing

In recent years,with the rapid development of tunnel construction in China,the mileage of tunnels has continuously increased,and the subsequent detection of tunnel diseases has received increasing attention from conservation departments.Among many diseases,the lining cracks are the most common and directly reflect the stress situation of the lining,which is very important for the study of tunnel diseases.At present,the detection of lining cracks depends on labor.This kind of observation by means of naked eyes and measurement by means of tools cannot guarantee the accuracy of the test results.In view of the current detection status and requirements,this paper conducts the research work on the vehicle-mounted tunnel lining crack detection system based on image processing.The main research contents are as follows:(1)In this paper,the function and overall structure of the system are studied.At the same time,the actual vehicle platform is built based on the design requirements and working principles of the system.The electric rail car,CCD camera,camera support,image acquisition card,strobe lighting device,rotation encoders and other equipment were selected and installed.In the actual detection process,the vehicle travels through the tunnel at a certain speed.Multiple cameras scan the lining surface to obtain crack images and save them in real time.Finally,crack information is obtained through offline processing of images.(2)In the process of preprocessing and stitching images,this paper uses the algorithm based on improved MSR and bilateral filtering fusion to solve the problem of non-uniform illumination of images.For the problem of the interference of the lining seam in the image,this paper uses the method based on the straight line and angle characteristics to remove it.Considering that the contrast between the crack and the lining background is not very obvious,the gray transformation and contrast enhancement were carried out to highlight the crack.At the same time,based on the SIFT scale invariant features and the weighted average fusion algorithm,the crack image sequence was stitched and fused.(3)In the process of identifying and classifying images,deriche filter are combined to perform edge detection on the image and an image segmentation method based on linear interpolation is used.The stitched panoramic lining image is divided into blocks,and maximum entropy is used for each block to calculate the global threshold.Then,the isolated noise is removed by morphological algorithm.The method is based on the shape feature of the crack region and the area extension method to remove the lumped noise and connect the fracture cracks.Finally,the morphological algorithm is adopted to refine the cracks and classify the extracted cracks according to the related characteristic parameters.(4)Based on the joint development environment of Halcon and Visual Studio,this system performs the detection experiment under the real tunnel scenes.The experimental results show that compared with the manual detection,the system detects the tunnel lining cracks quickly and accurately,which meets the requirements of the initial design.The research in this paper has achieved phased results and laid a solid foundation for the research projects that rely on it.