Boundary constrained voxel segmentation for 3D point clouds using local geometric differences

作者:

Highlights:

• A novel boundary control technique for segmentation method is proposed.

• An object segmentation benchmark dataset is prepared to segment different shapes.

• Two building segmentation benchmark datasets are prepared from outdoor scans.

• The proposed method gives better segmentation results than the other tested methods.

• The proposed method performs the process rather faster than the other methods.

摘要

•A novel boundary control technique for segmentation method is proposed.•An object segmentation benchmark dataset is prepared to segment different shapes.•Two building segmentation benchmark datasets are prepared from outdoor scans.•The proposed method gives better segmentation results than the other tested methods.•The proposed method performs the process rather faster than the other methods.

论文关键词:Octree,Point cloud segmentation,Segmentation dataset,Weighting

论文评审过程:Received 11 November 2019, Revised 5 April 2020, Accepted 5 April 2020, Available online 4 May 2020, Version of Record 13 May 2020.

论文官网地址:https://doi.org/10.1016/j.eswa.2020.113439