Deep Learning with WASI Simulation Data for Estimating Chlorophyll <i>a</i> Concentration of Inland Water Bodies
Information about the chlorophyll <i>a</i> concentration of inland water bodies is essential for water monitoring. This study focuses on estimating chlorophyll <i>a</i> with remote sensing data, and machine learning (ML) approaches on the real-world SpecWa dataset. We adapt a...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-02-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/4/718 |
| _version_ | 1797396413094559744 |
|---|---|
| author | Philipp M. Maier Sina Keller Stefan Hinz |
| author_facet | Philipp M. Maier Sina Keller Stefan Hinz |
| author_sort | Philipp M. Maier |
| collection | DOAJ |
| description | Information about the chlorophyll <i>a</i> concentration of inland water bodies is essential for water monitoring. This study focuses on estimating chlorophyll <i>a</i> with remote sensing data, and machine learning (ML) approaches on the real-world SpecWa dataset. We adapt and apply a one-dimensional convolutional neural network (1D CNN) as a deep learning architecture for the first time to address this estimation. Since such a DL approach requires a large amount of data for its training, we rely on simulation data generated by the Water Color Simulator (WASI). This simulation is prepared accordingly and includes a knowledge-based water composition with two origins of the chlorophyll <i>a</i> concentration. Therefore, the training data is independent of the real-world SpecWa dataset, which is challenging for any ML approach. We define two spectral downsampling approaches as a pre-processing step, representing the hyperspectral EnMAP satellite mission (<tt>SR-EnMAP</tt>) and the multispectral Sentinel-2 mission (<tt>SR-Sentinel</tt>). Subsequently, we train a Random Forest, an artificial neural network, a band-ratio approach, and the 1D CNN on the WASI-generated simulation training dataset. Finally, all ML models are evaluated on the real SpecWa dataset. For both downsampled data, the 1D CNN outperforms the other ML models. On the finer resolved <tt>SR-EnMAP</tt> data it achieves an <inline-formula><math display="inline"><semantics><mrow><msup><mi>R</mi><mn>2</mn></msup><mo>=</mo><mn>81.9</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math display="inline"><semantics><mrow><mi>R</mi><mi>M</mi><mi>S</mi><mi>E</mi><mo>=</mo><mn>12.4</mn></mrow></semantics></math></inline-formula> μg L<sup>−1</sup>, and <inline-formula><math display="inline"><semantics><mrow><mi>M</mi><mi>A</mi><mi>E</mi><mo>=</mo><mn>6.7</mn></mrow></semantics></math></inline-formula> μg L<sup>−1</sup>. Besides, the 1D CNN’s performance decreases on the <tt>SR-Sentinel</tt> data to <inline-formula><math display="inline"><semantics><mrow><msup><mi>R</mi><mn>2</mn></msup><mo>=</mo><mn>62.4</mn><mo>%</mo></mrow></semantics></math></inline-formula>. When focusing on the individual water bodies of the SpecWa dataset, the most significant differences exist between natural and artificial water bodies. We discover that the applied models estimate the chlorophyll <i>a</i> concentration of most natural water bodies satisfyingly. In sum, the newly DL approach can estimate the chlorophyll <i>a</i> values of unknown inland water bodies successfully, although it is trained on an entire simulation dataset. |
| first_indexed | 2024-03-09T00:49:56Z |
| format | Article |
| id | doaj.art-cf07a53db3824ca9b3758fa482248c4c |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T00:49:56Z |
| publishDate | 2021-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-cf07a53db3824ca9b3758fa482248c4c2023-12-11T17:16:21ZengMDPI AGRemote Sensing2072-42922021-02-0113471810.3390/rs13040718Deep Learning with WASI Simulation Data for Estimating Chlorophyll <i>a</i> Concentration of Inland Water BodiesPhilipp M. Maier0Sina Keller1Stefan Hinz2Institute of Photogrammetry and Remote Sensing, Karlsruhe Institute of Technology, 76131 Karlsruhe, GermanyInstitute of Photogrammetry and Remote Sensing, Karlsruhe Institute of Technology, 76131 Karlsruhe, GermanyInstitute of Photogrammetry and Remote Sensing, Karlsruhe Institute of Technology, 76131 Karlsruhe, GermanyInformation about the chlorophyll <i>a</i> concentration of inland water bodies is essential for water monitoring. This study focuses on estimating chlorophyll <i>a</i> with remote sensing data, and machine learning (ML) approaches on the real-world SpecWa dataset. We adapt and apply a one-dimensional convolutional neural network (1D CNN) as a deep learning architecture for the first time to address this estimation. Since such a DL approach requires a large amount of data for its training, we rely on simulation data generated by the Water Color Simulator (WASI). This simulation is prepared accordingly and includes a knowledge-based water composition with two origins of the chlorophyll <i>a</i> concentration. Therefore, the training data is independent of the real-world SpecWa dataset, which is challenging for any ML approach. We define two spectral downsampling approaches as a pre-processing step, representing the hyperspectral EnMAP satellite mission (<tt>SR-EnMAP</tt>) and the multispectral Sentinel-2 mission (<tt>SR-Sentinel</tt>). Subsequently, we train a Random Forest, an artificial neural network, a band-ratio approach, and the 1D CNN on the WASI-generated simulation training dataset. Finally, all ML models are evaluated on the real SpecWa dataset. For both downsampled data, the 1D CNN outperforms the other ML models. On the finer resolved <tt>SR-EnMAP</tt> data it achieves an <inline-formula><math display="inline"><semantics><mrow><msup><mi>R</mi><mn>2</mn></msup><mo>=</mo><mn>81.9</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math display="inline"><semantics><mrow><mi>R</mi><mi>M</mi><mi>S</mi><mi>E</mi><mo>=</mo><mn>12.4</mn></mrow></semantics></math></inline-formula> μg L<sup>−1</sup>, and <inline-formula><math display="inline"><semantics><mrow><mi>M</mi><mi>A</mi><mi>E</mi><mo>=</mo><mn>6.7</mn></mrow></semantics></math></inline-formula> μg L<sup>−1</sup>. Besides, the 1D CNN’s performance decreases on the <tt>SR-Sentinel</tt> data to <inline-formula><math display="inline"><semantics><mrow><msup><mi>R</mi><mn>2</mn></msup><mo>=</mo><mn>62.4</mn><mo>%</mo></mrow></semantics></math></inline-formula>. When focusing on the individual water bodies of the SpecWa dataset, the most significant differences exist between natural and artificial water bodies. We discover that the applied models estimate the chlorophyll <i>a</i> concentration of most natural water bodies satisfyingly. In sum, the newly DL approach can estimate the chlorophyll <i>a</i> values of unknown inland water bodies successfully, although it is trained on an entire simulation dataset.https://www.mdpi.com/2072-4292/13/4/718machine learningregressionCNNartificial neural networkradiative transfer modelWASI |
| spellingShingle | Philipp M. Maier Sina Keller Stefan Hinz Deep Learning with WASI Simulation Data for Estimating Chlorophyll <i>a</i> Concentration of Inland Water Bodies Remote Sensing machine learning regression CNN artificial neural network radiative transfer model WASI |
| title | Deep Learning with WASI Simulation Data for Estimating Chlorophyll <i>a</i> Concentration of Inland Water Bodies |
| title_full | Deep Learning with WASI Simulation Data for Estimating Chlorophyll <i>a</i> Concentration of Inland Water Bodies |
| title_fullStr | Deep Learning with WASI Simulation Data for Estimating Chlorophyll <i>a</i> Concentration of Inland Water Bodies |
| title_full_unstemmed | Deep Learning with WASI Simulation Data for Estimating Chlorophyll <i>a</i> Concentration of Inland Water Bodies |
| title_short | Deep Learning with WASI Simulation Data for Estimating Chlorophyll <i>a</i> Concentration of Inland Water Bodies |
| title_sort | deep learning with wasi simulation data for estimating chlorophyll i a i concentration of inland water bodies |
| topic | machine learning regression CNN artificial neural network radiative transfer model WASI |
| url | https://www.mdpi.com/2072-4292/13/4/718 |
| work_keys_str_mv | AT philippmmaier deeplearningwithwasisimulationdataforestimatingchlorophylliaiconcentrationofinlandwaterbodies AT sinakeller deeplearningwithwasisimulationdataforestimatingchlorophylliaiconcentrationofinlandwaterbodies AT stefanhinz deeplearningwithwasisimulationdataforestimatingchlorophylliaiconcentrationofinlandwaterbodies |