Pavement Crack Extraction Method from Mobile Laser Scanning Point Cloud

Abstract

The automatic detection of cracks in the road surface is of great significance for maintaining the road surface to ensure the safety of moving vehicles. The Hokuyo UTM 30 LX 2D laser scanner in this study is used to observe the road surface containing two cracks and distress areas. As a result, dense point clouds are created. The road surface is automatically extracted from the point clouds based on the geometry of the two curb lines. The normal vector of the points is calculated based on the principal component analysis method. Points belonging to cracks and zones of distress are extracted from the intensity gradient and the inclination angle between the two normal vectors of the neighboring points and then converted to binary images. The crack edges are extracted based on the Sobel operator. Although salt-and-pepper spots due to crack points extraction using intensity gradient affect the definition of crack edges, especially small cracks, large cracks and distress areas are extracted clearly. Research results show that reflectance intensity and elevation variation combination lead to the efficiency of crack extraction and distress area.

Visual reconstruction method of architectural space under laser point cloud big data

“…In this paper, facing the actual needs of 3D reconstruction of building space, aiming at the shortcomings of 3D point cloud data generation, processing and model generation, the 3D reconstruction of building is deeply studied. Visual reconstruction of architectural space has become a key issue in the field of computer vision [9], which has the same high application value as face recognition and target tracking. Three-dimensional reconstruction was first established on the basis of binocular vision, and then developed into multi-view reconstruction.…”

Pavement crack detection based on point cloud data and data fusion

“…It takes full advantage of efficient machine vision algorithms and avoids complex three-dimensional calculations. However, those approaches miss some edge details of pavement cracks [32][33][34][35][36]. Therefore, some research directly used three-dimensional data exaction for crack detection and demonstrated some successes [31,37,38].…”