Tunnel Crack Image Acquisition And Detection Technology Based On Area Scan Camera

As one of the main diseases of subway tunnel lining,crack has a serious negative impact on the safe operation of subway.Emerging technologies based on machine vision and image processing is gradually playing an important role in tunnel crack detection.This paper proposes a tunnel image acquisition system based on area-scan camera which can realize the continuous image acquisition of tunnel,and contrasts this system with line-scan camera system for the purpose of verifying which camera has better acquisition quality.A tunnel crack detection algorithm is designed and presented through an interface,which can realize identification and labeling of tunnel cracks.In addition,a three-dimensional display algorithm is designed to provide a visual demonstration of tunnel internal environment.In this paper,tunnel crack detection system based on area-scan camera is built that includes image acquisition system and image processing system.The image acquisition system which consists of camera,light source,computer and power supply,captures tunnel images and sends them to the computer for crack detection.A series of image processing operations are necessary for crack recognition due to the uneven quality of captured tunnel images.Image preprocessing is performed for the purpose of enhancing the image quality and acquiring binary image with cracks and interferents.Image denoising is used to filter out different kinds of interferents from the binary image,and then crack thinning and burr removal are used to get a clean crack skeleton map which is used in width calculation via the criterion.The crack area and its width data are labeled in the original image.In order to display the internal environment of the tunnel visually,this paper proposes a 3D display technology for subway tunnels.Implementing a 3D display of a tunnel requires two core steps:3D modeling and texture mapping.According to practical tunnel environment,a 3D tunnel model is established,and tunnel panoramic image is texture-mapped on the model to realize the tunnel 3D display.The image acquisition experiments of two systems are carried out in the laboratory and Guangzhou Metro respectively,and captured images are processed through the purposed interface.The experimental results show that images captured by line-scan camera are of good effect on continuity,but they are easily affected by the fixed structure of camera,which leads negative effects such as jitter and deformation.Images captured by area-scan camera are stable,but there are many overlapping areas among the images,which leads to a great amount of data.After image processing,the crack recognition effect of area-scan camera is obviously better than that of line-scan camera in the case of approximate image processing speed.In conclusion,image acquisition and processing technology based on area-scan camera has a solid theoretical basis and excellent application prospects in the field of tunnel crack detection.