An Accurate and Efficient Supervoxel Re-Segmentation Approach for Large-Scale Point Clouds Using Plane Constraints

The accurate and efficient segmentation of large-scale urban point clouds is crucial for many higher-level tasks, such as boundary line extraction, cloud registration, deformation measurement. In this paper, we propose a novel supervoxel approach to address the problem under-segmentation in local regions at various resolutions. Our introduces distance constraints from points planes merging stage enhance accuracy between non-coplanar supervoxels. Additionally, supervoxels with roughness above threshold are re-segmented using random sample consensus (RANSAC) multi-planar coupling within areas clouds. We tested proposed method on two publicly available datasets. results show that new outperforms classical methods terms recall, error, average entropy scenes.