A feature perturbation weakly supervised learning network for airborne multispectral LiDAR pointcloud classification
Currently, most pointcloud classification methods heavily rely on huge numbers of labeled samples. Notably, labeling a large-scale multispectral LiDAR (MS-LiDAR) pointcloud is time-consuming and costly. To address this issue, we propose a feature perturbation weakly supervised network for classifyin...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-03-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843224000372 |
| _version_ | 1797303380342734848 |
|---|---|
| author | Ke Chen Haiyan Guan Lanying Wang Yongtao Yu Yufu Zang Nannan Qin Jiacheng Liu Jonathan Li |
| author_facet | Ke Chen Haiyan Guan Lanying Wang Yongtao Yu Yufu Zang Nannan Qin Jiacheng Liu Jonathan Li |
| author_sort | Ke Chen |
| collection | DOAJ |
| description | Currently, most pointcloud classification methods heavily rely on huge numbers of labeled samples. Notably, labeling a large-scale multispectral LiDAR (MS-LiDAR) pointcloud is time-consuming and costly. To address this issue, we propose a feature perturbation weakly supervised network for classifying MS-LiDAR pointclouds using a few labeled samples, termed as FPWS-Net. In the FPWS-Net, we innovatively design a dual semantic inference structure, including a primary semantic inference module and an auxiliary semantic inference module. To provide the network with rich, accurate supervised signals, we embed kernel point convolution (KPConv) into the network for modelling the contextual information of MS-LiDAR pointclouds and propagating the signals between labeled and unlabeled points. Additionally, to constrain feature perturbations resulting from the dual semantic inference structure, fully leverage unlabeled points, and fit the architecture of the FPWS-Net, we combine consistency constraint and mutual pseudo-labeling loss. The proposed FPWS-Net is tested on two datasets, and achieves at least an average F1-score of 83.69 %, an mIoU of 78.81 %, and an OA of 95.97 % using only 0.1 % labeled points. The comparative experimental results demonstrate that the FPWS-Net not only outperforms the state-of-the-art (SOTA) weakly supervised networks, but also achieves the comparable classification performance to the fully supervised methods in the airborne MS-LiDAR pointcloud classification tasks. |
| first_indexed | 2024-03-07T23:51:54Z |
| format | Article |
| id | doaj.art-4b76c57983b94889a64742e336b70e4e |
| institution | Directory Open Access Journal |
| issn | 1569-8432 |
| language | English |
| last_indexed | 2024-03-07T23:51:54Z |
| publishDate | 2024-03-01 |
| publisher | Elsevier |
| record_format | Article |
| series | International Journal of Applied Earth Observations and Geoinformation |
| spelling | doaj.art-4b76c57983b94889a64742e336b70e4e2024-02-19T04:13:16ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322024-03-01127103683A feature perturbation weakly supervised learning network for airborne multispectral LiDAR pointcloud classificationKe Chen0Haiyan Guan1Lanying Wang2Yongtao Yu3Yufu Zang4Nannan Qin5Jiacheng Liu6Jonathan Li7School of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, China; Corresponding author.Department of Geography and Environmental Management, University of Waterloo, Waterloo ON N2L 3G1, CanadaFaculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huaian 223003, ChinaSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaDepartment of Geography and Environmental Management, University of Waterloo, Waterloo ON N2L 3G1, Canada; Department of Systems Design Engineering, University of Waterloo, Waterloo ON N2L 3G1, CanadaCurrently, most pointcloud classification methods heavily rely on huge numbers of labeled samples. Notably, labeling a large-scale multispectral LiDAR (MS-LiDAR) pointcloud is time-consuming and costly. To address this issue, we propose a feature perturbation weakly supervised network for classifying MS-LiDAR pointclouds using a few labeled samples, termed as FPWS-Net. In the FPWS-Net, we innovatively design a dual semantic inference structure, including a primary semantic inference module and an auxiliary semantic inference module. To provide the network with rich, accurate supervised signals, we embed kernel point convolution (KPConv) into the network for modelling the contextual information of MS-LiDAR pointclouds and propagating the signals between labeled and unlabeled points. Additionally, to constrain feature perturbations resulting from the dual semantic inference structure, fully leverage unlabeled points, and fit the architecture of the FPWS-Net, we combine consistency constraint and mutual pseudo-labeling loss. The proposed FPWS-Net is tested on two datasets, and achieves at least an average F1-score of 83.69 %, an mIoU of 78.81 %, and an OA of 95.97 % using only 0.1 % labeled points. The comparative experimental results demonstrate that the FPWS-Net not only outperforms the state-of-the-art (SOTA) weakly supervised networks, but also achieves the comparable classification performance to the fully supervised methods in the airborne MS-LiDAR pointcloud classification tasks.http://www.sciencedirect.com/science/article/pii/S1569843224000372Multispectral LiDARWeakly supervised learningPointcloud classificationDual semantic inference structureConsistency constraintMutual pseudo-labeling loss |
| spellingShingle | Ke Chen Haiyan Guan Lanying Wang Yongtao Yu Yufu Zang Nannan Qin Jiacheng Liu Jonathan Li A feature perturbation weakly supervised learning network for airborne multispectral LiDAR pointcloud classification International Journal of Applied Earth Observations and Geoinformation Multispectral LiDAR Weakly supervised learning Pointcloud classification Dual semantic inference structure Consistency constraint Mutual pseudo-labeling loss |
| title | A feature perturbation weakly supervised learning network for airborne multispectral LiDAR pointcloud classification |
| title_full | A feature perturbation weakly supervised learning network for airborne multispectral LiDAR pointcloud classification |
| title_fullStr | A feature perturbation weakly supervised learning network for airborne multispectral LiDAR pointcloud classification |
| title_full_unstemmed | A feature perturbation weakly supervised learning network for airborne multispectral LiDAR pointcloud classification |
| title_short | A feature perturbation weakly supervised learning network for airborne multispectral LiDAR pointcloud classification |
| title_sort | feature perturbation weakly supervised learning network for airborne multispectral lidar pointcloud classification |
| topic | Multispectral LiDAR Weakly supervised learning Pointcloud classification Dual semantic inference structure Consistency constraint Mutual pseudo-labeling loss |
| url | http://www.sciencedirect.com/science/article/pii/S1569843224000372 |
| work_keys_str_mv | AT kechen afeatureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT haiyanguan afeatureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT lanyingwang afeatureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT yongtaoyu afeatureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT yufuzang afeatureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT nannanqin afeatureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT jiachengliu afeatureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT jonathanli afeatureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT kechen featureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT haiyanguan featureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT lanyingwang featureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT yongtaoyu featureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT yufuzang featureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT nannanqin featureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT jiachengliu featureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification AT jonathanli featureperturbationweaklysupervisedlearningnetworkforairbornemultispectrallidarpointcloudclassification |