Self-Supervised Learning for Precise Individual Tree Segmentation in Airborne LiDAR Point Clouds
Segmenting individual trees from airborne LiDAR point cloud data is critical for forest management, urban planning, and ecological monitoring but remains challenging due to complex natural environments, diverse tree architectures, and dense canopies. Traditional supervised methods rely on extensive,...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10973613/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850170892500336640 |
|---|---|
| author | Lama Shaheen Bader Rasheed Manuel Mazzara |
| author_facet | Lama Shaheen Bader Rasheed Manuel Mazzara |
| author_sort | Lama Shaheen |
| collection | DOAJ |
| description | Segmenting individual trees from airborne LiDAR point cloud data is critical for forest management, urban planning, and ecological monitoring but remains challenging due to complex natural environments, diverse tree architectures, and dense canopies. Traditional supervised methods rely on extensive, manually annotated datasets that are often impractical to obtain. In this study, we propose a novel self-supervised learning framework that eliminates the need for manual labeling by integrating transformation-invariant feature extraction, an energy-based segmentation loss, and soft clustering. The framework operates in two stages: a pretext task applies geometric transformations—rotation (from –45° to +45°), translation (between –1 and 1 units), and scaling (between 0.5 and 2.0)—to learn robust features, while an unsupervised segmentation step leverages an energy function that combines height, density, and slope attributes to cluster points corresponding to individual trees. We evaluated our approach on a high-density LiDAR dataset acquired from the Estonian Land Board (LAS format, version 1.4) comprising over 850,000 points. Our method achieves up to 87% convexity, 78% solidity, and an elliptical fit error as low as 0.12, substantially reducing over-segmentation compared to baseline clustering techniques. These results demonstrate that our self-supervised framework offers a scalable, label-free solution for precise tree segmentation, with significant advantages in accuracy and efficiency over traditional supervised methods. |
| format | Article |
| id | doaj-art-8f75db01cc294f13a74909c41ef62b0c |
| institution | OA Journals |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-8f75db01cc294f13a74909c41ef62b0c2025-08-20T02:20:23ZengIEEEIEEE Access2169-35362025-01-0113708957090810.1109/ACCESS.2025.356336310973613Self-Supervised Learning for Precise Individual Tree Segmentation in Airborne LiDAR Point CloudsLama Shaheen0Bader Rasheed1https://orcid.org/0000-0003-3874-0883Manuel Mazzara2https://orcid.org/0000-0002-3860-4948Institute of Software Development and Engineering, Innopolis University, Innopolis, RussiaLaboratory of Innovative Technologies for Processing Video Content, Innopolis University, Innopolis, RussiaInstitute of Software Development and Engineering, Innopolis University, Innopolis, RussiaSegmenting individual trees from airborne LiDAR point cloud data is critical for forest management, urban planning, and ecological monitoring but remains challenging due to complex natural environments, diverse tree architectures, and dense canopies. Traditional supervised methods rely on extensive, manually annotated datasets that are often impractical to obtain. In this study, we propose a novel self-supervised learning framework that eliminates the need for manual labeling by integrating transformation-invariant feature extraction, an energy-based segmentation loss, and soft clustering. The framework operates in two stages: a pretext task applies geometric transformations—rotation (from –45° to +45°), translation (between –1 and 1 units), and scaling (between 0.5 and 2.0)—to learn robust features, while an unsupervised segmentation step leverages an energy function that combines height, density, and slope attributes to cluster points corresponding to individual trees. We evaluated our approach on a high-density LiDAR dataset acquired from the Estonian Land Board (LAS format, version 1.4) comprising over 850,000 points. Our method achieves up to 87% convexity, 78% solidity, and an elliptical fit error as low as 0.12, substantially reducing over-segmentation compared to baseline clustering techniques. These results demonstrate that our self-supervised framework offers a scalable, label-free solution for precise tree segmentation, with significant advantages in accuracy and efficiency over traditional supervised methods.https://ieeexplore.ieee.org/document/10973613/Self-supervised learningindividual trees segmentationairborne LiDAR data processing |
| spellingShingle | Lama Shaheen Bader Rasheed Manuel Mazzara Self-Supervised Learning for Precise Individual Tree Segmentation in Airborne LiDAR Point Clouds IEEE Access Self-supervised learning individual trees segmentation airborne LiDAR data processing |
| title | Self-Supervised Learning for Precise Individual Tree Segmentation in Airborne LiDAR Point Clouds |
| title_full | Self-Supervised Learning for Precise Individual Tree Segmentation in Airborne LiDAR Point Clouds |
| title_fullStr | Self-Supervised Learning for Precise Individual Tree Segmentation in Airborne LiDAR Point Clouds |
| title_full_unstemmed | Self-Supervised Learning for Precise Individual Tree Segmentation in Airborne LiDAR Point Clouds |
| title_short | Self-Supervised Learning for Precise Individual Tree Segmentation in Airborne LiDAR Point Clouds |
| title_sort | self supervised learning for precise individual tree segmentation in airborne lidar point clouds |
| topic | Self-supervised learning individual trees segmentation airborne LiDAR data processing |
| url | https://ieeexplore.ieee.org/document/10973613/ |
| work_keys_str_mv | AT lamashaheen selfsupervisedlearningforpreciseindividualtreesegmentationinairbornelidarpointclouds AT baderrasheed selfsupervisedlearningforpreciseindividualtreesegmentationinairbornelidarpointclouds AT manuelmazzara selfsupervisedlearningforpreciseindividualtreesegmentationinairbornelidarpointclouds |