Learning Digital Terrain Models From Point Clouds: ALS2DTM Dataset and Rasterization-Based GAN
Despite the popularity of deep neural networks in various domains, the extraction of digital terrain models (DTMs) from airborne laser scanning (ALS) point clouds is still challenging. This might be due to the lack of the dedicated large-scale annotated dataset and the data-structure discrepancy bet...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9794452/ |
| _version_ | 1811332036962549760 |
|---|---|
| author | Hoang-An Le Florent Guiotte Minh-Tan Pham Sebastien Lefevre Thomas Corpetti |
| author_facet | Hoang-An Le Florent Guiotte Minh-Tan Pham Sebastien Lefevre Thomas Corpetti |
| author_sort | Hoang-An Le |
| collection | DOAJ |
| description | Despite the popularity of deep neural networks in various domains, the extraction of digital terrain models (DTMs) from airborne laser scanning (ALS) point clouds is still challenging. This might be due to the lack of the dedicated large-scale annotated dataset and the data-structure discrepancy between point clouds and DTMs. To promote data-driven DTM extraction, this article collects from open sources a large-scale dataset of ALS point clouds and corresponding DTMs with various urban, forested, and mountainous scenes. A baseline method is proposed as the first attempt to train a deep neural network to extract DTMs directly from ALS point clouds via rasterization techniques, coined DeepTerRa. Extensive studies with well-established methods are performed to benchmark the dataset and analyze the challenges in learning to extract DTM from point clouds. The experimental results show the interest of the agnostic data-driven approach, with submetric error level compared to methods designed for DTM extraction. The data and source code are available online at <uri>https://lhoangan.github.io/deepterra/</uri> for reproducibility and further similar research. |
| first_indexed | 2024-04-13T16:29:45Z |
| format | Article |
| id | doaj.art-8eca5ead4b5e40a89e4f6171a8d4705e |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-04-13T16:29:45Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-8eca5ead4b5e40a89e4f6171a8d4705e2022-12-22T02:39:37ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352022-01-01154980498910.1109/JSTARS.2022.31820309794452Learning Digital Terrain Models From Point Clouds: ALS2DTM Dataset and Rasterization-Based GANHoang-An Le0https://orcid.org/0000-0002-7896-5967Florent Guiotte1Minh-Tan Pham2https://orcid.org/0000-0003-0266-767XSebastien Lefevre3https://orcid.org/0000-0002-2384-8202Thomas Corpetti4Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université Bretagne Sud, Vannes, FranceInstitut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université Bretagne Sud, Vannes, FranceInstitut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université Bretagne Sud, Vannes, FranceInstitut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université Bretagne Sud, Vannes, FranceLittoral - Environnement - Télédétection - Géomatique (LETG), UMR6554, Université Rennes 2, Rennes, FranceDespite the popularity of deep neural networks in various domains, the extraction of digital terrain models (DTMs) from airborne laser scanning (ALS) point clouds is still challenging. This might be due to the lack of the dedicated large-scale annotated dataset and the data-structure discrepancy between point clouds and DTMs. To promote data-driven DTM extraction, this article collects from open sources a large-scale dataset of ALS point clouds and corresponding DTMs with various urban, forested, and mountainous scenes. A baseline method is proposed as the first attempt to train a deep neural network to extract DTMs directly from ALS point clouds via rasterization techniques, coined DeepTerRa. Extensive studies with well-established methods are performed to benchmark the dataset and analyze the challenges in learning to extract DTM from point clouds. The experimental results show the interest of the agnostic data-driven approach, with submetric error level compared to methods designed for DTM extraction. The data and source code are available online at <uri>https://lhoangan.github.io/deepterra/</uri> for reproducibility and further similar research.https://ieeexplore.ieee.org/document/9794452/Airborne laser scanning (ALS) point clouddatasetdeep networksdigital terrain model (DTM)generative adversarial network (GAN)rasterization |
| spellingShingle | Hoang-An Le Florent Guiotte Minh-Tan Pham Sebastien Lefevre Thomas Corpetti Learning Digital Terrain Models From Point Clouds: ALS2DTM Dataset and Rasterization-Based GAN IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Airborne laser scanning (ALS) point cloud dataset deep networks digital terrain model (DTM) generative adversarial network (GAN) rasterization |
| title | Learning Digital Terrain Models From Point Clouds: ALS2DTM Dataset and Rasterization-Based GAN |
| title_full | Learning Digital Terrain Models From Point Clouds: ALS2DTM Dataset and Rasterization-Based GAN |
| title_fullStr | Learning Digital Terrain Models From Point Clouds: ALS2DTM Dataset and Rasterization-Based GAN |
| title_full_unstemmed | Learning Digital Terrain Models From Point Clouds: ALS2DTM Dataset and Rasterization-Based GAN |
| title_short | Learning Digital Terrain Models From Point Clouds: ALS2DTM Dataset and Rasterization-Based GAN |
| title_sort | learning digital terrain models from point clouds als2dtm dataset and rasterization based gan |
| topic | Airborne laser scanning (ALS) point cloud dataset deep networks digital terrain model (DTM) generative adversarial network (GAN) rasterization |
| url | https://ieeexplore.ieee.org/document/9794452/ |
| work_keys_str_mv | AT hoanganle learningdigitalterrainmodelsfrompointcloudsals2dtmdatasetandrasterizationbasedgan AT florentguiotte learningdigitalterrainmodelsfrompointcloudsals2dtmdatasetandrasterizationbasedgan AT minhtanpham learningdigitalterrainmodelsfrompointcloudsals2dtmdatasetandrasterizationbasedgan AT sebastienlefevre learningdigitalterrainmodelsfrompointcloudsals2dtmdatasetandrasterizationbasedgan AT thomascorpetti learningdigitalterrainmodelsfrompointcloudsals2dtmdatasetandrasterizationbasedgan |