Coarse‐to‐fine adjustment for multi‐platform point cloud fusion
Xin Zhao,Jianping Li,Yuhao Li,Bisheng Yang,Sihan Sun,Yongfeng Lin,Zhen Dong
DOI: https://doi.org/10.1111/phor.12513
2024-07-26
The Photogrammetric Record
Abstract:For the four types of point cloud data (TLS, ALS, MLS and BPLS), first, the point clouds from the mobile platform (ALS, MLS and BPLS) are segmented according to trajectory information. Then, bounding boxes are established for all point cloud blocks, and smoothness‐type constraint edges and registration‐type constraint edges are constructed using trajectory and bounding box information. Subsequently, various feature points are extracted and selected from the point cloud blocks with registration‐type constraint edges to obtain corresponding point pairs. Using corresponding point pairs, a multi‐scale error equation is constructed and solved, providing initial values for global optimisation. Finally, error equations are constructed for raw point correspondences, and iterative solution is performed in a tightly coupled manner to achieve global optimisation. Leveraging multi‐platform laser scanning systems offers a complete solution for 3D modelling of large‐scale urban scenes. However, the spatial inconsistency of point clouds collected by heterogeneous platforms with different viewpoints presents challenges in achieving seamless fusion. To tackle this challenge, this paper proposes a coarse‐to‐fine adjustment for multi‐platform point cloud fusion. First, in the preprocessing stage, the bounding box of each point cloud block is employed to identify potential constraint association. Second, the proposed local optimisation facilitates preliminary pairwise alignment with these potential constraint relationships, and obtaining initial guess for a comprehensive global optimisation. At last, the proposed global optimisation incorporates all the local constraints for tightly coupled optimisation with raw point correspondences. We choose two study areas to conduct experiments. Study area 1 represents a fast road scene with a significant amount of vegetation, while study area 2 represents an urban scene with many buildings. Extensive experimental evaluations indicate the proposed method has increased the accuracy of study area 1 by 50.6% and the accuracy of study area 2 by 44.7%.
geosciences, multidisciplinary,geography, physical,remote sensing,imaging science & photographic technology