The 20th and 21st joint COTA International Conference of Transportation Professionals (CICTP 2020-21)

Due to the poor safety and low efficiency of the traditional road pothole detection method, a road pothole detection method based on a 3D laser point cloud is proposed. Firstly, based on the Pauta criterion, the abnormal points are detected, and the left or right normal points are used to correct the abnormal points; Secondly, establish a quadratic parabolic arched pavement model, and judge whether there are potholes in a cross-section based on the dispersion of the point cloud data distribution of the cross section; Subsequently, a method of twice fitting of arch pavement curve is proposed to extract pothole points using the differences in data before and after fitting; Finally, based on the detection accuracy of the point cloud data in the cross-section and vertical-sections, the minimum area and volume detected at each data point is calculated to give information on the area and volume of the entire pothole. Experiments have proved that the method can accurately extract the pothole points, and the calculated pothole data has higher consistency compared with manual data.

CICTP