Predicting Vegetation Stratum Occupancy from Airborne LiDAR Data with Deep Learning
We propose a new deep learning-based method for estimating the occupancy of vegetation strata from airborne 3D LiDAR point clouds. Our model predicts rasterized occupancy maps for three vegetation strata corresponding to lower, medium, and higher cover. Our weakly-supervised training scheme allows o...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-08-01
|
| Series: | International Journal of Applied Earth Observations and Geoinformation |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1569843222000656 |
| _version_ | 1817998677194571776 |
|---|---|
| author | Ekaterina Kalinicheva Loic Landrieu Clément Mallet Nesrine Chehata |
| author_facet | Ekaterina Kalinicheva Loic Landrieu Clément Mallet Nesrine Chehata |
| author_sort | Ekaterina Kalinicheva |
| collection | DOAJ |
| description | We propose a new deep learning-based method for estimating the occupancy of vegetation strata from airborne 3D LiDAR point clouds. Our model predicts rasterized occupancy maps for three vegetation strata corresponding to lower, medium, and higher cover. Our weakly-supervised training scheme allows our network to only be supervised with vegetation occupancy values aggregated over cylindrical plots containing thousands of points which are typically easier to produce than pixel-wise or point-wise annotations. We propose to employ a deep neural network operating on 3D points, and whose prediction are projected onto rasters representing the different vegetation strata. Our method outperforms handcrafted, regression and deep learning baselines in terms of precision by up to 30%, while simultaneously providing visual and interpretable predictions. We provide an open-source implementation along with a dataset of 199 agricultural plots to train and evaluate weakly supervised occupancy regression algorithms. |
| first_indexed | 2024-04-14T02:56:37Z |
| format | Article |
| id | doaj.art-1c256d016c3a46ae9eb4a0e4e8ef952f |
| institution | Directory Open Access Journal |
| issn | 1569-8432 |
| language | English |
| last_indexed | 2024-04-14T02:56:37Z |
| publishDate | 2022-08-01 |
| publisher | Elsevier |
| record_format | Article |
| series | International Journal of Applied Earth Observations and Geoinformation |
| spelling | doaj.art-1c256d016c3a46ae9eb4a0e4e8ef952f2022-12-22T02:16:03ZengElsevierInternational Journal of Applied Earth Observations and Geoinformation1569-84322022-08-01112102863Predicting Vegetation Stratum Occupancy from Airborne LiDAR Data with Deep LearningEkaterina Kalinicheva0Loic Landrieu1Clément Mallet2Nesrine Chehata3LASTIG, Univ Gustave Eiffel, IGN, ENSG, F-94160 Saint-Mande, France; INRAE, UMR 1202 BIOGECO, Université de Bordeaux, FranceLASTIG, Univ Gustave Eiffel, IGN, ENSG, F-94160 Saint-Mande, FranceLASTIG, Univ Gustave Eiffel, IGN, ENSG, F-94160 Saint-Mande, FranceLASTIG, Univ Gustave Eiffel, IGN, ENSG, F-94160 Saint-Mande, France; EA G&E Bordeaux INP, Université Bordeaux Montaigne, FranceWe propose a new deep learning-based method for estimating the occupancy of vegetation strata from airborne 3D LiDAR point clouds. Our model predicts rasterized occupancy maps for three vegetation strata corresponding to lower, medium, and higher cover. Our weakly-supervised training scheme allows our network to only be supervised with vegetation occupancy values aggregated over cylindrical plots containing thousands of points which are typically easier to produce than pixel-wise or point-wise annotations. We propose to employ a deep neural network operating on 3D points, and whose prediction are projected onto rasters representing the different vegetation strata. Our method outperforms handcrafted, regression and deep learning baselines in terms of precision by up to 30%, while simultaneously providing visual and interpretable predictions. We provide an open-source implementation along with a dataset of 199 agricultural plots to train and evaluate weakly supervised occupancy regression algorithms.http://www.sciencedirect.com/science/article/pii/S1569843222000656Vegetation analysisStratumAirborne LiDARDeep neural networksWeakly-supervised learning |
| spellingShingle | Ekaterina Kalinicheva Loic Landrieu Clément Mallet Nesrine Chehata Predicting Vegetation Stratum Occupancy from Airborne LiDAR Data with Deep Learning International Journal of Applied Earth Observations and Geoinformation Vegetation analysis Stratum Airborne LiDAR Deep neural networks Weakly-supervised learning |
| title | Predicting Vegetation Stratum Occupancy from Airborne LiDAR Data with Deep Learning |
| title_full | Predicting Vegetation Stratum Occupancy from Airborne LiDAR Data with Deep Learning |
| title_fullStr | Predicting Vegetation Stratum Occupancy from Airborne LiDAR Data with Deep Learning |
| title_full_unstemmed | Predicting Vegetation Stratum Occupancy from Airborne LiDAR Data with Deep Learning |
| title_short | Predicting Vegetation Stratum Occupancy from Airborne LiDAR Data with Deep Learning |
| title_sort | predicting vegetation stratum occupancy from airborne lidar data with deep learning |
| topic | Vegetation analysis Stratum Airborne LiDAR Deep neural networks Weakly-supervised learning |
| url | http://www.sciencedirect.com/science/article/pii/S1569843222000656 |
| work_keys_str_mv | AT ekaterinakalinicheva predictingvegetationstratumoccupancyfromairbornelidardatawithdeeplearning AT loiclandrieu predictingvegetationstratumoccupancyfromairbornelidardatawithdeeplearning AT clementmallet predictingvegetationstratumoccupancyfromairbornelidardatawithdeeplearning AT nesrinechehata predictingvegetationstratumoccupancyfromairbornelidardatawithdeeplearning |