Normal estimation is a crucial first step for numerous light detection and ranging (LiDAR) data-processing algorithms, from building reconstruction, road extraction, and ground-cover classification to scene rendering. For LiDAR point clouds in urban environments, this paper presents a robust method to estimate normals by constructing an octree-based hierarchical representation for the data and detecting a group of large enough consistent neighborhoods at multiscales. Consistent neighborhoods are mainly determined based on the observation that an urban environment is typically comprised of regular objects, e.g., buildings, roads, and the ground surface, and irregular objects, e.g., trees and shrubs; the surfaces of most regular objects can be approximatively represented by a group of local planes. Even in the frequent presence of heavy noise and anisotropic point samplings in LiDAR data, our method is capable of estimating robust normals for kinds of objects in urban environments, and the estimated normals are beneficial to more accurately segment and identify the objects, as well as preserving their sharp features and complete outlines. The proposed method was experimentally validated both on synthetic and real urban LiDAR datasets, and was compared to state-of-the-art methods.

Light detection and ranging (LiDAR) scanning has become a prevalent technique for digitalizing outdoor scenes with three-dimensional (3D) point clouds. An automatic segmentation of the LiDAR data is important for understanding and reconstructing the outdoor scenes. However, it is still a challenging task due to complex and various objects in the outdoor scenes and some scanning drawbacks in the LiDAR data. Observing that most of the objects, such as the ground, road, roof, and facade, can be locally described with a group of geometric shapes, e.g. plane, sphere, and cylinder, we propose an automatic method to segment raw LiDAR data by robustly extracting the shapes in this paper. Firstly, our method divides the raw LiDAR data into a number of supervoxels considering the geometric and spectral consistencies of LiDAR data. Secondly, we robustly extract shapes from each supervoxels using a random sample consensus (RANSAC)-based method and evaluates and optimises the extracted shapes with a density loss estimation technique. Finally, our method outputs the segmentation result by merging the extracted shapes into a group of complete shapes, each of them represents a meaningful object in the outdoor scene. Experiments show that the method is efficient and robust for extracting most of the shapes in the outdoor scenes from a given raw LiDAR point cloud, and no preprocessing is required.