A Multifeature Fusion Network for Tree Species Classification Based on Ground-Based LiDAR Data
Light detection and ranging (LiDAR) holds considerable promise for tree species classification. Existing networks that utilize point clouds of individual trees have shown promising results. However, challenges, such as incomplete point cloud data, uneven point density across different components of...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10834575/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1825207045215748096 |
|---|---|
| author | Yaoting Liu Yiming Chen Zhengjun Liu Jianchang Chen Yuxuan Liu |
| author_facet | Yaoting Liu Yiming Chen Zhengjun Liu Jianchang Chen Yuxuan Liu |
| author_sort | Yaoting Liu |
| collection | DOAJ |
| description | Light detection and ranging (LiDAR) holds considerable promise for tree species classification. Existing networks that utilize point clouds of individual trees have shown promising results. However, challenges, such as incomplete point cloud data, uneven point density across different components of the tree, and complex tree morphologies, can hinder classification accuracy. To overcome these limitations, we introduced the multifeature fusion tree classifier network (MFFTC-Net). This network leverages a novel boundary-driven point sampling method that preserves more canopy points and mitigates the effects of uneven point density. We also utilize the umbrella-repSurf module, which captures local geometric features and enhances the model's responsiveness to tree structural nuances. The backbone of MFFTC-Net integrates these innovations through a multifeature fusion approach, utilizing set abstraction for local information capture and transformer-based feature interaction for robust multiscale feature integration. Our results demonstrate that MFFTC-Net significantly outperforms other state-of-the-art methods in LiDAR-based tree species classification, achieving the highest overall accuracy and kappa coefficients on both a self-built dataset of four species and a public dataset of seven species. |
| format | Article |
| id | doaj-art-bbdbcb5102e343ccaede54ed1e540e0c |
| institution | Kabale University |
| issn | 1939-1404 2151-1535 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj-art-bbdbcb5102e343ccaede54ed1e540e0c2025-02-07T00:00:36ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01184648466310.1109/JSTARS.2025.352780810834575A Multifeature Fusion Network for Tree Species Classification Based on Ground-Based LiDAR DataYaoting Liu0https://orcid.org/0009-0009-6188-0818Yiming Chen1https://orcid.org/0009-0008-3856-1416Zhengjun Liu2https://orcid.org/0000-0002-0303-6290Jianchang Chen3Yuxuan Liu4https://orcid.org/0000-0003-4394-1989Chinese Academy of Surveying and Mapping, Beijing, ChinaChinese Academy of Surveying and Mapping, Beijing, ChinaChinese Academy of Surveying and Mapping, Beijing, ChinaWuhan University, Wuhan, ChinaChinese Academy of Surveying and Mapping, Beijing, ChinaLight detection and ranging (LiDAR) holds considerable promise for tree species classification. Existing networks that utilize point clouds of individual trees have shown promising results. However, challenges, such as incomplete point cloud data, uneven point density across different components of the tree, and complex tree morphologies, can hinder classification accuracy. To overcome these limitations, we introduced the multifeature fusion tree classifier network (MFFTC-Net). This network leverages a novel boundary-driven point sampling method that preserves more canopy points and mitigates the effects of uneven point density. We also utilize the umbrella-repSurf module, which captures local geometric features and enhances the model's responsiveness to tree structural nuances. The backbone of MFFTC-Net integrates these innovations through a multifeature fusion approach, utilizing set abstraction for local information capture and transformer-based feature interaction for robust multiscale feature integration. Our results demonstrate that MFFTC-Net significantly outperforms other state-of-the-art methods in LiDAR-based tree species classification, achieving the highest overall accuracy and kappa coefficients on both a self-built dataset of four species and a public dataset of seven species.https://ieeexplore.ieee.org/document/10834575/Deep learninglight detection and ranging (LiDAR)multifeature fusion tree classifier network (MFFTC-Net)tree species classification |
| spellingShingle | Yaoting Liu Yiming Chen Zhengjun Liu Jianchang Chen Yuxuan Liu A Multifeature Fusion Network for Tree Species Classification Based on Ground-Based LiDAR Data IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Deep learning light detection and ranging (LiDAR) multifeature fusion tree classifier network (MFFTC-Net) tree species classification |
| title | A Multifeature Fusion Network for Tree Species Classification Based on Ground-Based LiDAR Data |
| title_full | A Multifeature Fusion Network for Tree Species Classification Based on Ground-Based LiDAR Data |
| title_fullStr | A Multifeature Fusion Network for Tree Species Classification Based on Ground-Based LiDAR Data |
| title_full_unstemmed | A Multifeature Fusion Network for Tree Species Classification Based on Ground-Based LiDAR Data |
| title_short | A Multifeature Fusion Network for Tree Species Classification Based on Ground-Based LiDAR Data |
| title_sort | multifeature fusion network for tree species classification based on ground based lidar data |
| topic | Deep learning light detection and ranging (LiDAR) multifeature fusion tree classifier network (MFFTC-Net) tree species classification |
| url | https://ieeexplore.ieee.org/document/10834575/ |
| work_keys_str_mv | AT yaotingliu amultifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT yimingchen amultifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT zhengjunliu amultifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT jianchangchen amultifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT yuxuanliu amultifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT yaotingliu multifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT yimingchen multifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT zhengjunliu multifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT jianchangchen multifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata AT yuxuanliu multifeaturefusionnetworkfortreespeciesclassificationbasedongroundbasedlidardata |