PCRMLP: A Two-Stage Network for Point Cloud Registration in Urban Scenes
Point cloud registration plays a crucial role in 3D mapping and localization. Urban scene point clouds pose significant challenges for registration due to their large data volume, similar scenarios, and dynamic objects. Estimating the location by instances (bulidings, traffic lights, etc.) in urban...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-06-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/12/5758 |
| _version_ | 1797592638697766912 |
|---|---|
| author | Jingyang Liu Yucheng Xu Lu Zhou Lei Sun |
| author_facet | Jingyang Liu Yucheng Xu Lu Zhou Lei Sun |
| author_sort | Jingyang Liu |
| collection | DOAJ |
| description | Point cloud registration plays a crucial role in 3D mapping and localization. Urban scene point clouds pose significant challenges for registration due to their large data volume, similar scenarios, and dynamic objects. Estimating the location by instances (bulidings, traffic lights, etc.) in urban scenes is a more humanized matter. In this paper, we propose PCRMLP (point cloud registration MLP), a novel model for urban scene point cloud registration that achieves comparable registration performance to prior learning-based methods. Compared to previous works that focused on extracting features and estimating correspondence, PCRMLP estimates transformation implicitly from concrete instances. The key innovation lies in the instance-level urban scene representation method, which leverages semantic segmentation and density-based spatial clustering of applications with noise (DBSCAN) to generate instance descriptors, enabling robust feature extraction, dynamic object filtering, and logical transformation estimation. Then, a lightweight network consisting of Multilayer Perceptrons (MLPs) is employed to obtain transformation in an encoder–decoder manner. Experimental validation on the KITTI dataset demonstrates that PCRMLP achieves satisfactory coarse transformation estimates from instance descriptors within a remarkable time of 0.0028 s. With the incorporation of an ICP refinement module, our proposed method outperforms prior learning-based approaches, yielding a rotation error of 2.01° and a translation error of 1.58 m. The experimental results highlight PCRMLP’s potential for coarse registration of urban scene point clouds, thereby paving the way for its application in instance-level semantic mapping and localization. |
| first_indexed | 2024-03-11T01:55:54Z |
| format | Article |
| id | doaj.art-608723ede5854fa4a363a5f6f0b3a851 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T01:55:54Z |
| publishDate | 2023-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-608723ede5854fa4a363a5f6f0b3a8512023-11-18T12:35:52ZengMDPI AGSensors1424-82202023-06-012312575810.3390/s23125758PCRMLP: A Two-Stage Network for Point Cloud Registration in Urban ScenesJingyang Liu0Yucheng Xu1Lu Zhou2Lei Sun3College of Artificial Intelligence, Nankai University, Tianjin 300071, ChinaSchool of Informatics, University of Edinburgh, Edinburgh EH8 9YL, UKCollege of Artificial Intelligence, Nankai University, Tianjin 300071, ChinaCollege of Artificial Intelligence, Nankai University, Tianjin 300071, ChinaPoint cloud registration plays a crucial role in 3D mapping and localization. Urban scene point clouds pose significant challenges for registration due to their large data volume, similar scenarios, and dynamic objects. Estimating the location by instances (bulidings, traffic lights, etc.) in urban scenes is a more humanized matter. In this paper, we propose PCRMLP (point cloud registration MLP), a novel model for urban scene point cloud registration that achieves comparable registration performance to prior learning-based methods. Compared to previous works that focused on extracting features and estimating correspondence, PCRMLP estimates transformation implicitly from concrete instances. The key innovation lies in the instance-level urban scene representation method, which leverages semantic segmentation and density-based spatial clustering of applications with noise (DBSCAN) to generate instance descriptors, enabling robust feature extraction, dynamic object filtering, and logical transformation estimation. Then, a lightweight network consisting of Multilayer Perceptrons (MLPs) is employed to obtain transformation in an encoder–decoder manner. Experimental validation on the KITTI dataset demonstrates that PCRMLP achieves satisfactory coarse transformation estimates from instance descriptors within a remarkable time of 0.0028 s. With the incorporation of an ICP refinement module, our proposed method outperforms prior learning-based approaches, yielding a rotation error of 2.01° and a translation error of 1.58 m. The experimental results highlight PCRMLP’s potential for coarse registration of urban scene point clouds, thereby paving the way for its application in instance-level semantic mapping and localization.https://www.mdpi.com/1424-8220/23/12/5758DBSCANdeep learninginstance levelpoint cloudsregistrationurban scene |
| spellingShingle | Jingyang Liu Yucheng Xu Lu Zhou Lei Sun PCRMLP: A Two-Stage Network for Point Cloud Registration in Urban Scenes Sensors DBSCAN deep learning instance level point clouds registration urban scene |
| title | PCRMLP: A Two-Stage Network for Point Cloud Registration in Urban Scenes |
| title_full | PCRMLP: A Two-Stage Network for Point Cloud Registration in Urban Scenes |
| title_fullStr | PCRMLP: A Two-Stage Network for Point Cloud Registration in Urban Scenes |
| title_full_unstemmed | PCRMLP: A Two-Stage Network for Point Cloud Registration in Urban Scenes |
| title_short | PCRMLP: A Two-Stage Network for Point Cloud Registration in Urban Scenes |
| title_sort | pcrmlp a two stage network for point cloud registration in urban scenes |
| topic | DBSCAN deep learning instance level point clouds registration urban scene |
| url | https://www.mdpi.com/1424-8220/23/12/5758 |
| work_keys_str_mv | AT jingyangliu pcrmlpatwostagenetworkforpointcloudregistrationinurbanscenes AT yuchengxu pcrmlpatwostagenetworkforpointcloudregistrationinurbanscenes AT luzhou pcrmlpatwostagenetworkforpointcloudregistrationinurbanscenes AT leisun pcrmlpatwostagenetworkforpointcloudregistrationinurbanscenes |