Research And Design Of Tunnel Lining Crack Detection System Based On Image Processing

With the acceleration of the national modernization process, the demand oftransportation for quick increase year by year, therefore, the number of thetunnel increases. Tunnel disease problem along with tunnel construction and thedisease is complex and changeable, if not to detect of tunnel diseases timely, alarge number of resources and money will be lost, thus detection of tunneldiseases is necessary, even a must. Cracks is one of the most common diseasesin tunnel, also the source of other diseases, if not test the crack in time, it willdeform, seriously affect the safe use of the tunnel, also a threat to the person andour property, Based on these factors, cracks in tunnel must be detected, crackshere is to point to the surface of the tunnel lining cracks. At present theapplication of automation testing equipment in the tunnel in our country is stillin the primary stage, the application is not wide, detection of the tunnel ismainly by manual detection, but the manual detection needs long time, highlabor intensity, also it has a certain degree of risk, so researching on theautomatic tunnel crack detection system is meaningful.This paper firstly introduces the caused reasons, types and harms about thecracks in tunnel lining, gives the evaluation standard of tunnel cracks, elaboratessome methods at home and abroad in the detection of cracks in tunnel, analyzessome problems encountered in the detection of cracks in tunnel. Then this paperintroduces the detection system, the system is divided into two parts: image acquisition system and image processing system. Image acquisition systemmainly includes the carrier vehicle, CCD camera, capture card, lightingequipment, positioning system and image processing system is mainly toprocess and analyze the collected images.Next, according to the characteristics of tunnel lining crack image collected,processing and analysis will be done in the pictures. First of all, due to thecamera itself, the aberration of the collected image is an inevitable problem, thedistorted image will be corrected. Unlike the pavement, tunnel exists curvedsurface, so the pixel on the curved surface needs to be corrected, these steps willmake a convenient to the later parameter calculation. In order to obtain the rapidand accurate detection of crack image, this paper has had an intensiveprocessing to the collected image will be performed. Compare severalenhancement algorithms, Retinex algorithm based on the Guided filter for imageenhancement is adopted. The experimental results show that the algorithm has agood effect on the image enhancement. Further, compare image somesegmentation algorithm, image segmentation is conducted based on edgedetection. After image segmentation, crack area separates from the backgroundregion, and then in this paper, the cracks of the image are classified. Accordingto the projective characteristics of the pixel in cracks, crack image is dividedinto horizontal cracks, vertical cracks and reticular cracks, and the parametersfor crack images are calculated, the calculated parameters value of cracksprovides a scientific and reliable basis for the evaluation of tunnel cracks.