Research Of Pavement Crack Automatic Detection Technology Based On The Stereo Photogrammetry

With the rapid growth of road construction, the pavement quality inspection andmaintenance work has become increasingly important that attach great importance of thetraffic management departments. Crack of pavement is one of the main diseases. How todetect cracks in pavement damage efficiently, objectively and accurately has a very importantsignificance and application value on China’s road maintenance management.This paper introduces the damaged road surface crack detection and identificationtechnologies and systems development status in detail. Putting forward a kind of automaticdetection technology based on three-dimensional structure of the cross after analyzing thecharacteristics and technical difficulties of pavement crack detection technology. This methodcollects pavement cracks information from different angles. This technology gains moreinformation compared with the monocular shooting in order to detect crack more simply andaccurately. This paper mainly did the following aspects of work for the research of automaticcrack detection technology road damage based on stereoscopic photography.1. This paper put forward stereo binocular camera crack detection technology afteranalyzing the existing deficiencies and difficulties of current monocular camera crackdetection technology and the ways to the problem that monocular camera can not solve. Andit describes the basic principle of stereo-photography and analyzes stereo vision systemarchitecture and gives a measure based on non-stereoscopic shooting calculated pavementcracks namely eikonometer relationship equation of points, surface, substance.2. Designs the overall layout of the structure of damaged pavement crack detectionsystem based on stereo photogrammetric respectively from the model test andvehicle-mounted two aspects and introduces the principle and method of selecting the mainhardware devices.3. Research the image processing techniques on damaged pavement crack detection. Inthe image preprocessing operations, firstly, Equalization of image processing. Secondly,eliminate image noise through the median filter. Convert the image and then computing a newimage from the converted image and the original image, further binarizing the new image. Finally, mathematical morphology is applied to the binary image in order to remove glitches,holes, cracks connectivity and the ideal binary crack image and the crack length informationare appearing.4. Introducing the calculation method of the three-dimensional shoot camera calibrationand coefficient and carried out the camera calibration experiments. Getting the thesisexperimental parameters through pavement cracks stereoscopic shooting test systemcalibration experiments.By the above studies, it verifies the system which is based on stereo photogrammetricdetection can efficiently, objectively, accurately detect the pavement damage.