Rethinking of learning-based 3D keypoints detection for large-scale point clouds registration
The main solution for large-scale point clouds registration is to first obtain a set of matched 3D keypoint pairs and then accomplish the point cloud registration task based on these matched keypoint pairs. However, at present, many methods study the feature descriptors in the point clouds registrat...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-08-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843222001418 |
| _version_ | 1817998651211907072 |
|---|---|
| author | ShaoCong Liu Tao Wang Yan Zhang Ruqin Zhou Chenguang Dai Yongsheng Zhang Haozhen Lei Hanyun Wang |
| author_facet | ShaoCong Liu Tao Wang Yan Zhang Ruqin Zhou Chenguang Dai Yongsheng Zhang Haozhen Lei Hanyun Wang |
| author_sort | ShaoCong Liu |
| collection | DOAJ |
| description | The main solution for large-scale point clouds registration is to first obtain a set of matched 3D keypoint pairs and then accomplish the point cloud registration task based on these matched keypoint pairs. However, at present, many methods study the feature descriptors in the point clouds registration task, but few methods discuss the 3D keypoints detection issues. The commonly used 3D keypoints detection strategy is the voxel-grid-based downsampling method, and the detected 3D keypoints are usually with a relatively huge amount and also with no explicit geometrical properties, which finally leads to a low inlier ratio. In this study, we rethink the 3D keypoints detection problem for large-scale point clouds with deep learning. Specifically, we discuss four kinds of 3D keypoints detection methods based on the joint keypoint detection and description learning framework D3Feat, and carry out extensive analyses on both the indoor large-scale point clouds dataset 3DMatch and the outdoor large-scale point clouds dataset KITTI Odometry. Experimental results demonstrate that the Multi-layer Perceptron (MLP) based method achieves the best inlier ratios under the different numbers of extracted 3D keypoints on both the indoor and outdoor large-scale point clouds. Further, we test these four kinds of keypoints detection methods under the application of large-scale point clouds registration, and the MLP-based method also achieves the state-of-the-art registration performance. |
| first_indexed | 2024-04-14T02:55:41Z |
| format | Article |
| id | doaj.art-5d012de997e24270ba753dd2519e05e0 |
| institution | Directory Open Access Journal |
| issn | 1569-8432 |
| language | English |
| last_indexed | 2024-04-14T02:55:41Z |
| publishDate | 2022-08-01 |
| publisher | Elsevier |
| record_format | Article |
| series | International Journal of Applied Earth Observations and Geoinformation |
| spelling | doaj.art-5d012de997e24270ba753dd2519e05e02022-12-22T02:16:06ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322022-08-01112102944Rethinking of learning-based 3D keypoints detection for large-scale point clouds registrationShaoCong Liu0Tao Wang1Yan Zhang2Ruqin Zhou3Chenguang Dai4Yongsheng Zhang5Haozhen Lei6Hanyun Wang7School of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, ChinaSchool of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, ChinaSchool of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, ChinaSchool of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, ChinaSchool of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, ChinaSchool of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, ChinaSchool of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, ChinaSchool of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, China; State Key Laboratory of Resources and Environmental Information System, Beijing 100000, China; Corresponding author at: School of Surveying and Mapping, Information Engineering University, Zhengzhou 450001, China.The main solution for large-scale point clouds registration is to first obtain a set of matched 3D keypoint pairs and then accomplish the point cloud registration task based on these matched keypoint pairs. However, at present, many methods study the feature descriptors in the point clouds registration task, but few methods discuss the 3D keypoints detection issues. The commonly used 3D keypoints detection strategy is the voxel-grid-based downsampling method, and the detected 3D keypoints are usually with a relatively huge amount and also with no explicit geometrical properties, which finally leads to a low inlier ratio. In this study, we rethink the 3D keypoints detection problem for large-scale point clouds with deep learning. Specifically, we discuss four kinds of 3D keypoints detection methods based on the joint keypoint detection and description learning framework D3Feat, and carry out extensive analyses on both the indoor large-scale point clouds dataset 3DMatch and the outdoor large-scale point clouds dataset KITTI Odometry. Experimental results demonstrate that the Multi-layer Perceptron (MLP) based method achieves the best inlier ratios under the different numbers of extracted 3D keypoints on both the indoor and outdoor large-scale point clouds. Further, we test these four kinds of keypoints detection methods under the application of large-scale point clouds registration, and the MLP-based method also achieves the state-of-the-art registration performance.http://www.sciencedirect.com/science/article/pii/S1569843222001418Large-scale point clouds3D keypoints detectionDeep learningPoint clouds registration |
| spellingShingle | ShaoCong Liu Tao Wang Yan Zhang Ruqin Zhou Chenguang Dai Yongsheng Zhang Haozhen Lei Hanyun Wang Rethinking of learning-based 3D keypoints detection for large-scale point clouds registration International Journal of Applied Earth Observations and Geoinformation Large-scale point clouds 3D keypoints detection Deep learning Point clouds registration |
| title | Rethinking of learning-based 3D keypoints detection for large-scale point clouds registration |
| title_full | Rethinking of learning-based 3D keypoints detection for large-scale point clouds registration |
| title_fullStr | Rethinking of learning-based 3D keypoints detection for large-scale point clouds registration |
| title_full_unstemmed | Rethinking of learning-based 3D keypoints detection for large-scale point clouds registration |
| title_short | Rethinking of learning-based 3D keypoints detection for large-scale point clouds registration |
| title_sort | rethinking of learning based 3d keypoints detection for large scale point clouds registration |
| topic | Large-scale point clouds 3D keypoints detection Deep learning Point clouds registration |
| url | http://www.sciencedirect.com/science/article/pii/S1569843222001418 |
| work_keys_str_mv | AT shaocongliu rethinkingoflearningbased3dkeypointsdetectionforlargescalepointcloudsregistration AT taowang rethinkingoflearningbased3dkeypointsdetectionforlargescalepointcloudsregistration AT yanzhang rethinkingoflearningbased3dkeypointsdetectionforlargescalepointcloudsregistration AT ruqinzhou rethinkingoflearningbased3dkeypointsdetectionforlargescalepointcloudsregistration AT chenguangdai rethinkingoflearningbased3dkeypointsdetectionforlargescalepointcloudsregistration AT yongshengzhang rethinkingoflearningbased3dkeypointsdetectionforlargescalepointcloudsregistration AT haozhenlei rethinkingoflearningbased3dkeypointsdetectionforlargescalepointcloudsregistration AT hanyunwang rethinkingoflearningbased3dkeypointsdetectionforlargescalepointcloudsregistration |