Abstract:Vehicle-borne laser scanning (VLS) and backpack laser scanning (BLS) are two emerging mobile mapping technologies for capturing detailed spatial information near ground in urban built environments. BLS has flexible mobility and usually provides point clouds in a local coordinate system. Therefore, a mismatch between VLS and BLS point cloud data is quite common. Fusing VLS and BLS data in different coordinate systems could provide a comprehensive survey of urban built environments. Because of the complexity of urban road environments and the difference in data acquisition methods, traditional registration approaches based on point-level correspondences are likely to fail and sometimes involve substantial manual efforts. In this article, we propose a novel registration approach that finds the optimal transformation between the respective point clouds based on a unique tree distribution pattern defined by tree trunk centers. The proposed method consists of three key steps, i.e., trunk center extraction, stepwise minimum spanning tree (MST) matching, and transformation estimation. Stepwise MST matching is an essential step in finding the one-to-one correspondences using a topological similarity between the two light detection and ranging (LiDAR) datasets. We evaluated our method with five real-world datasets collected in Shanghai, China. The results showed that the proposed method performed well in all five experiment sites with an average rotation error of less than 0.06° and an average translation error of less than 0.05 m. Moreover, the reported mean position deviation in the five sites is 0.112, 0.144, 0.176, 0.148, and 0.184 m. Our proposed method has a great potential for registering multiplatform LiDAR data that could provide comprehensive and essential 3-D information for numerous urban applications.

Local to Global Registration of Multi-View Range Scans Using Spanning Tree.

Range Image Registration Using Hierarchical Segmentation and Clustering

Automatic multi-view registration of unordered range scans without feature extraction

Multi-view registration of unordered range scans by fast correspondence propagation of multi-scale descriptors

Multi-view Non-rigid Registration Based on Multiple Thin-plate Splines

Scaling Registration Of Multiview Range Scans Via Motion Averaging

A New Method for Registration of 3d Point Sets with Low Overlapping Ratios

Hierarchical Multiview Rigid Registration

Template-Guided Hierarchical Multi-View Registration Framework of Unordered Bridge Terrestrial Laser Scanning Data

Surface reconstruction via efficient and accurate registration of multiview range scans

Registration of Multiple Laser Scans Based on 3D Contour Features.

Globally Consistent Rigid Registration.

Multi-View Laser Point Cloud Global Registration for a Single Object

Study on Multi-Views Point Clouds Registration

Registration of Multiple Range Data from Real Scenes Using 3D Contour Features

Minimum Spanning Tree Hierarchically Fusing Multi-feature Points and High-Dimensional Features for Medical Image Registration.

Minimum Spanning Tree Fusing Multi-Salient Points Hierarchically For Multi-Modality Image Registration

A Stepwise Minimum Spanning Tree Matching Method for Registering Vehicle-Borne and Backpack LiDAR Point Clouds

Automatic global registration of multiple 3D data sets from outdoor urban environments based on feature units.

Fully automated structured light scanning for high-fidelity 3D reconstruction via graph optimization

An integration method for scanned multi-view range images (MRIs) based on local weighted least squares (LWLS) surface fitting