Automatic registration of tree point clouds from terrestrial laser scanning

Multiple scans are generally required to fully reconstruct three-dimensional models of botanical trees. An algorithm for the automatic registration of tree point clouds from terrestrial laser scanning is proposed in this letter. The method extracts skeletons from the point clouds and conducts skeleton-based registration automatically. It uses a closed-form solution for the fine registration of the skeletons. An ICP procedure is applied to the roughly aligned tree point clouds to refine registration results of the skeleton-based method. Two example trees are registered using the proposed algorithm. The algorithm does not require a perfect skeleton to be extracted. No manual coarse registration is needed. The algorithm runs slower than our previously proposed method but with better registration accuracy. The algorithm contributes to the automatic marker-free tree point cloud registration and improves the registration accuracy of existing methods.