Terrain-Net: A Highly-Efficient, Parameter-Free, and Easy-to-Use Deep Neural Network for Ground Filtering of UAV LiDAR Data in Forested Environments
In recent years, a rise in interest in using Unmanned Aerial Vehicles (UAV) with LiDAR (Light Detection and Ranging) to capture the 3D structure of forests for forestry and ecosystem monitoring applications has been witnessed. Since the terrain is an essential basis for the vertical structure modeli...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-11-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/22/5798 |
| _version_ | 1797464056697716736 |
|---|---|
| author | Bowen Li Hao Lu Han Wang Jianbo Qi Gang Yang Yong Pang Haolin Dong Yining Lian |
| author_facet | Bowen Li Hao Lu Han Wang Jianbo Qi Gang Yang Yong Pang Haolin Dong Yining Lian |
| author_sort | Bowen Li |
| collection | DOAJ |
| description | In recent years, a rise in interest in using Unmanned Aerial Vehicles (UAV) with LiDAR (Light Detection and Ranging) to capture the 3D structure of forests for forestry and ecosystem monitoring applications has been witnessed. Since the terrain is an essential basis for the vertical structure modeling of a forest, the point cloud filtering delivering a highly accurate Digital Terrain Model (DTM) contributes significantly to forest studies. Conventional point cloud filtering algorithms require users to select suitable parameters according to the knowledge of the algorithm and the characteristics of scanned scenes, which are normally empirical and time-consuming. Deep learning offers a novel method in classifying and segmenting LiDAR point cloud, while there are only few studies reported on utilizing deep learning to filter non-ground LiDAR points of forested environments. In this study, we proposed an end-to-end and highly-efficient network named Terrain-net which combines the 3D point convolution operator and self-attention mechanism to capture local and global features for UAV point cloud ground filtering. The network was trained with over 15 million labeled points of 70 forest sites and was evaluated at 17 sites covering various forested environments. Terrain-net was compared with four classical filtering algorithms and one of the most well-recognized point convolution-based deep learning methods (KP-FCNN). Results indicated that Terrain-net achieved the best performance in respect of the Kappa coefficient (0.93), MIoU (0.933) and overall accuracy (98.0%). Terrain-net also performed well in transferring to an additional third-party open dataset for ground filtering in large-scale scenes and other vegetated environments. No parameters need to be tuned in transferring predictions. Terrain-net will hopefully be widely applied as a new highly-efficient, parameter-free, and easy-to-use tool for LiDAR data ground filtering in varying forest environments. |
| first_indexed | 2024-03-09T18:02:30Z |
| format | Article |
| id | doaj.art-ef0778a27dcf48adbf4676b290597c72 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T18:02:30Z |
| publishDate | 2022-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-ef0778a27dcf48adbf4676b290597c722023-11-24T09:50:36ZengMDPI AGRemote Sensing2072-42922022-11-011422579810.3390/rs14225798Terrain-Net: A Highly-Efficient, Parameter-Free, and Easy-to-Use Deep Neural Network for Ground Filtering of UAV LiDAR Data in Forested EnvironmentsBowen Li0Hao Lu1Han Wang2Jianbo Qi3Gang Yang4Yong Pang5Haolin Dong6Yining Lian7School of Information Science and Technology, Beijing Forestry University, Beijing 100083, ChinaSchool of Information Science and Technology, Beijing Forestry University, Beijing 100083, ChinaSchool of Information Science and Technology, Beijing Forestry University, Beijing 100083, ChinaResearch Center of Forest Management Engineering of State Forestry and Grassland Administration, Beijing Forestry University, Beijing 100083, ChinaSchool of Information Science and Technology, Beijing Forestry University, Beijing 100083, ChinaInstitute of Forest Resource Information Techniques, Chinese Academy of Forestry, Beijing 100091, ChinaSchool of Information Science and Technology, Beijing Forestry University, Beijing 100083, ChinaSchool of Information Science and Technology, Beijing Forestry University, Beijing 100083, ChinaIn recent years, a rise in interest in using Unmanned Aerial Vehicles (UAV) with LiDAR (Light Detection and Ranging) to capture the 3D structure of forests for forestry and ecosystem monitoring applications has been witnessed. Since the terrain is an essential basis for the vertical structure modeling of a forest, the point cloud filtering delivering a highly accurate Digital Terrain Model (DTM) contributes significantly to forest studies. Conventional point cloud filtering algorithms require users to select suitable parameters according to the knowledge of the algorithm and the characteristics of scanned scenes, which are normally empirical and time-consuming. Deep learning offers a novel method in classifying and segmenting LiDAR point cloud, while there are only few studies reported on utilizing deep learning to filter non-ground LiDAR points of forested environments. In this study, we proposed an end-to-end and highly-efficient network named Terrain-net which combines the 3D point convolution operator and self-attention mechanism to capture local and global features for UAV point cloud ground filtering. The network was trained with over 15 million labeled points of 70 forest sites and was evaluated at 17 sites covering various forested environments. Terrain-net was compared with four classical filtering algorithms and one of the most well-recognized point convolution-based deep learning methods (KP-FCNN). Results indicated that Terrain-net achieved the best performance in respect of the Kappa coefficient (0.93), MIoU (0.933) and overall accuracy (98.0%). Terrain-net also performed well in transferring to an additional third-party open dataset for ground filtering in large-scale scenes and other vegetated environments. No parameters need to be tuned in transferring predictions. Terrain-net will hopefully be widely applied as a new highly-efficient, parameter-free, and easy-to-use tool for LiDAR data ground filtering in varying forest environments.https://www.mdpi.com/2072-4292/14/22/5798UAVLiDARground filteringdeep learningforestry |
| spellingShingle | Bowen Li Hao Lu Han Wang Jianbo Qi Gang Yang Yong Pang Haolin Dong Yining Lian Terrain-Net: A Highly-Efficient, Parameter-Free, and Easy-to-Use Deep Neural Network for Ground Filtering of UAV LiDAR Data in Forested Environments Remote Sensing UAV LiDAR ground filtering deep learning forestry |
| title | Terrain-Net: A Highly-Efficient, Parameter-Free, and Easy-to-Use Deep Neural Network for Ground Filtering of UAV LiDAR Data in Forested Environments |
| title_full | Terrain-Net: A Highly-Efficient, Parameter-Free, and Easy-to-Use Deep Neural Network for Ground Filtering of UAV LiDAR Data in Forested Environments |
| title_fullStr | Terrain-Net: A Highly-Efficient, Parameter-Free, and Easy-to-Use Deep Neural Network for Ground Filtering of UAV LiDAR Data in Forested Environments |
| title_full_unstemmed | Terrain-Net: A Highly-Efficient, Parameter-Free, and Easy-to-Use Deep Neural Network for Ground Filtering of UAV LiDAR Data in Forested Environments |
| title_short | Terrain-Net: A Highly-Efficient, Parameter-Free, and Easy-to-Use Deep Neural Network for Ground Filtering of UAV LiDAR Data in Forested Environments |
| title_sort | terrain net a highly efficient parameter free and easy to use deep neural network for ground filtering of uav lidar data in forested environments |
| topic | UAV LiDAR ground filtering deep learning forestry |
| url | https://www.mdpi.com/2072-4292/14/22/5798 |
| work_keys_str_mv | AT bowenli terrainnetahighlyefficientparameterfreeandeasytousedeepneuralnetworkforgroundfilteringofuavlidardatainforestedenvironments AT haolu terrainnetahighlyefficientparameterfreeandeasytousedeepneuralnetworkforgroundfilteringofuavlidardatainforestedenvironments AT hanwang terrainnetahighlyefficientparameterfreeandeasytousedeepneuralnetworkforgroundfilteringofuavlidardatainforestedenvironments AT jianboqi terrainnetahighlyefficientparameterfreeandeasytousedeepneuralnetworkforgroundfilteringofuavlidardatainforestedenvironments AT gangyang terrainnetahighlyefficientparameterfreeandeasytousedeepneuralnetworkforgroundfilteringofuavlidardatainforestedenvironments AT yongpang terrainnetahighlyefficientparameterfreeandeasytousedeepneuralnetworkforgroundfilteringofuavlidardatainforestedenvironments AT haolindong terrainnetahighlyefficientparameterfreeandeasytousedeepneuralnetworkforgroundfilteringofuavlidardatainforestedenvironments AT yininglian terrainnetahighlyefficientparameterfreeandeasytousedeepneuralnetworkforgroundfilteringofuavlidardatainforestedenvironments |