Detection Of Airport Runway Pavement Crack Based On Image Processing Technology

The roughness of airport runway surface is a very important indicator of flight safety to flightdepartments. Any tiny flaw of the runway surface would be impacted when the airplanes roll and takeoff. And it would occur flying accident if there was a little uneven on the pavement of airportrunway.The detection and recognition on the pavement crack of the airport runway was studied in thisthesis. The purpose of the paper is to judge the pavement whether there is a crack in the detection andrecognition process by the image pretreatment, image segmentation and mathematics morphologicalanalysis of the airport runway pavement. If there is a crack, then extract the crack’s shape, andcalculate the crack’s length, width and other geometric characteristics. The main tasks of this thesis areas follows:1. An improved algorithm for the weighted mean filter was proposed in this thesis. Theambiguity of image and the image processing fuzzy degree could be reduced better than typicalfiltering algorithms.2. An edge detection algorithm on wavelet-based was proposed in this thesis. In order to achievethe purpose of effective separation of the cracks, the characteristics of cracks were analyzed. Severalcommonly used edge detection algorithms were compared. The iteration segmentation was selectedafter comparing their advantages and disadvantages, separating the background and cracks in thepavement while retains the original crack shape information.3. An algorithm of extracting geometric feature for the pavement cracks of the airport fieldrunway was proposed. The feature of the image and the processing on morphology was extracted. Thescattered points and the burrs of the image through the method of closing and opening were processedin this thesis. Then the skeleton of the image was extracted, the length and width of the crack wascalculated. The shape of the crack information was extracted better than the proposed algorithm.The results showed that the runway pavement crack detection and recognition were bettercompleted and achieved the desired goal, which provided some basic work for the airport runwaygradual automation of routine maintenance.