Study On Pavement Pothole Detection Technology Based On Video3-dimensional Reconstruction

The detection of pavement damage is the most important part of expressway maintenance and management. As the mileage of expressways increasing continuously, the traditional manual ways of detecting can not meet the requirements of speed and accuracy any longer. In recent years, the electrooptical technology and computer technology are rapidly developing, and the technical specifications of electronics hardware are continuously improving, which makes automate detection of pavement possible. As a result, the pavement damage detection system based on video detection technology has become a hotspot of current research. This article has designed a pavement pothole detection system based on video detection technology, which can be used to detect the depth and width of the pothole and reconstruct it in a three dimensional way, and also has had a further research on the algorithms related to the detection process of the system.Firstly, this article introduces the composition of the system and the principle of detection based on LED line light source. Then we calibrate the CCD camera which takes images for the system. The article introduces a few kinds of common ways of calibrating cameras, and chooses to calibrate the camera with RAC two-step method from traditional calibration according to the merit and demerit of all those ways, taking the specific circumstance of the experiment into consideration.The gathered images would be preconditioned by image processing module, including image gray processing, image enhancement, median filtering and image binaryzation, to get rid of image noise and make the key information of the image prominent. This article introduces the principle of edge detection, part of the image segmentation technology, and the edge of the characteristic circle from the calibration board is extracted. Then the central point of target image from the light band is extracted by mathematical morphological processing.After camera calibration and image processing, the section of the pothole is reconstructed according to the corresponding relation between the coordinates, and the error rang between calculated value and observed value of the width and depth of the pothole is figured out. By intersecting two mutually perpendicular light bands, the three dimensional coordinates can be figured out, and the three dimensional image of the pothole can be reconstructed with VRML, which is also introduced in the paper. Finally, the model is improved according to the analysis about the error rang of the calibration board characteristic point extracting and light bands image gathering, both of which are typical errors of the experiment. This detection system can get considerable accurate information about the depth and width of the pothole, and the three dimensional algorithm has a simple principle and faster calculating speed, which makes it possible to reconstruct a major three dimensional image of the pothole.