Extensibility of U-Net Neural Network Model for Hydrographic Feature Extraction and Implications for Hydrologic Modeling
Accurate maps of regional surface water features are integral for advancing ecologic, atmospheric and land development studies. The only comprehensive surface water feature map of Alaska is the National Hydrography Dataset (NHD). NHD features are often digitized representations of historic topograph...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-06-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/12/2368 |
| _version_ | 1827689504435077120 |
|---|---|
| author | Lawrence V. Stanislawski Ethan J. Shavers Shaowen Wang Zhe Jiang E. Lynn Usery Evan Moak Alexander Duffy Joel Schott |
| author_facet | Lawrence V. Stanislawski Ethan J. Shavers Shaowen Wang Zhe Jiang E. Lynn Usery Evan Moak Alexander Duffy Joel Schott |
| author_sort | Lawrence V. Stanislawski |
| collection | DOAJ |
| description | Accurate maps of regional surface water features are integral for advancing ecologic, atmospheric and land development studies. The only comprehensive surface water feature map of Alaska is the National Hydrography Dataset (NHD). NHD features are often digitized representations of historic topographic map blue lines and may be outdated. Here we test deep learning methods to automatically extract surface water features from airborne interferometric synthetic aperture radar (IfSAR) data to update and validate Alaska hydrographic databases. U-net artificial neural networks (ANN) and high-performance computing (HPC) are used for supervised hydrographic feature extraction within a study area comprised of 50 contiguous watersheds in Alaska. Surface water features derived from elevation through automated flow-routing and manual editing are used as training data. Model extensibility is tested with a series of 16 U-net models trained with increasing percentages of the study area, from about 3 to 35 percent. Hydrography is predicted by each of the models for all watersheds not used in training. Input raster layers are derived from digital terrain models, digital surface models, and intensity images from the IfSAR data. Results indicate about 15 percent of the study area is required to optimally train the ANN to extract hydrography when F1-scores for tested watersheds average between 66 and 68. Little benefit is gained by training beyond 15 percent of the study area. Fully connected hydrographic networks are generated for the U-net predictions using a novel approach that constrains a D-8 flow-routing approach to follow U-net predictions. This work demonstrates the ability of deep learning to derive surface water feature maps from complex terrain over a broad area. |
| first_indexed | 2024-03-10T10:19:36Z |
| format | Article |
| id | doaj.art-78e9c6102a4f4c19a8b30634d35a1f70 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T10:19:36Z |
| publishDate | 2021-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-78e9c6102a4f4c19a8b30634d35a1f702023-11-22T00:33:40ZengMDPI AGRemote Sensing2072-42922021-06-011312236810.3390/rs13122368Extensibility of U-Net Neural Network Model for Hydrographic Feature Extraction and Implications for Hydrologic ModelingLawrence V. Stanislawski0Ethan J. Shavers1Shaowen Wang2Zhe Jiang3E. Lynn Usery4Evan Moak5Alexander Duffy6Joel Schott7U.S. Geological Survey, Center of Excellence for Geospatial Information Science, Rolla, MO 65401, USAU.S. Geological Survey, Center of Excellence for Geospatial Information Science, Rolla, MO 65401, USADepartment of Geography and Geographic Information Science, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USADepartment of Computer & Information Science & Engineering, University of Florida, Gainesville, FL 32611, USAU.S. Geological Survey, Center of Excellence for Geospatial Information Science, Rolla, MO 65401, USACollege of Engineering and Computing, University of Missouri Science & Technology, Rolla, MO 65401, USACollege of Engineering and Computing, University of Missouri Science & Technology, Rolla, MO 65401, USACollege of Engineering and Computing, University of Missouri Science & Technology, Rolla, MO 65401, USAAccurate maps of regional surface water features are integral for advancing ecologic, atmospheric and land development studies. The only comprehensive surface water feature map of Alaska is the National Hydrography Dataset (NHD). NHD features are often digitized representations of historic topographic map blue lines and may be outdated. Here we test deep learning methods to automatically extract surface water features from airborne interferometric synthetic aperture radar (IfSAR) data to update and validate Alaska hydrographic databases. U-net artificial neural networks (ANN) and high-performance computing (HPC) are used for supervised hydrographic feature extraction within a study area comprised of 50 contiguous watersheds in Alaska. Surface water features derived from elevation through automated flow-routing and manual editing are used as training data. Model extensibility is tested with a series of 16 U-net models trained with increasing percentages of the study area, from about 3 to 35 percent. Hydrography is predicted by each of the models for all watersheds not used in training. Input raster layers are derived from digital terrain models, digital surface models, and intensity images from the IfSAR data. Results indicate about 15 percent of the study area is required to optimally train the ANN to extract hydrography when F1-scores for tested watersheds average between 66 and 68. Little benefit is gained by training beyond 15 percent of the study area. Fully connected hydrographic networks are generated for the U-net predictions using a novel approach that constrains a D-8 flow-routing approach to follow U-net predictions. This work demonstrates the ability of deep learning to derive surface water feature maps from complex terrain over a broad area.https://www.mdpi.com/2072-4292/13/12/2368machine learningneural networkU-netfeature extractionhydrography |
| spellingShingle | Lawrence V. Stanislawski Ethan J. Shavers Shaowen Wang Zhe Jiang E. Lynn Usery Evan Moak Alexander Duffy Joel Schott Extensibility of U-Net Neural Network Model for Hydrographic Feature Extraction and Implications for Hydrologic Modeling Remote Sensing machine learning neural network U-net feature extraction hydrography |
| title | Extensibility of U-Net Neural Network Model for Hydrographic Feature Extraction and Implications for Hydrologic Modeling |
| title_full | Extensibility of U-Net Neural Network Model for Hydrographic Feature Extraction and Implications for Hydrologic Modeling |
| title_fullStr | Extensibility of U-Net Neural Network Model for Hydrographic Feature Extraction and Implications for Hydrologic Modeling |
| title_full_unstemmed | Extensibility of U-Net Neural Network Model for Hydrographic Feature Extraction and Implications for Hydrologic Modeling |
| title_short | Extensibility of U-Net Neural Network Model for Hydrographic Feature Extraction and Implications for Hydrologic Modeling |
| title_sort | extensibility of u net neural network model for hydrographic feature extraction and implications for hydrologic modeling |
| topic | machine learning neural network U-net feature extraction hydrography |
| url | https://www.mdpi.com/2072-4292/13/12/2368 |
| work_keys_str_mv | AT lawrencevstanislawski extensibilityofunetneuralnetworkmodelforhydrographicfeatureextractionandimplicationsforhydrologicmodeling AT ethanjshavers extensibilityofunetneuralnetworkmodelforhydrographicfeatureextractionandimplicationsforhydrologicmodeling AT shaowenwang extensibilityofunetneuralnetworkmodelforhydrographicfeatureextractionandimplicationsforhydrologicmodeling AT zhejiang extensibilityofunetneuralnetworkmodelforhydrographicfeatureextractionandimplicationsforhydrologicmodeling AT elynnusery extensibilityofunetneuralnetworkmodelforhydrographicfeatureextractionandimplicationsforhydrologicmodeling AT evanmoak extensibilityofunetneuralnetworkmodelforhydrographicfeatureextractionandimplicationsforhydrologicmodeling AT alexanderduffy extensibilityofunetneuralnetworkmodelforhydrographicfeatureextractionandimplicationsforhydrologicmodeling AT joelschott extensibilityofunetneuralnetworkmodelforhydrographicfeatureextractionandimplicationsforhydrologicmodeling |