Advanced Fully Convolutional Networks for Agricultural Field Boundary Detection
Accurate spatial information of agricultural fields is important for providing actionable information to farmers, managers, and policymakers. On the other hand, the automated detection of field boundaries is a challenging task due to their small size, irregular shape and the use of mixed-cropping sy...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-02-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/13/4/722 |
_version_ | 1797396317038706688 |
---|---|
author | Alireza Taravat Matthias P. Wagner Rogerio Bonifacio David Petit |
author_facet | Alireza Taravat Matthias P. Wagner Rogerio Bonifacio David Petit |
author_sort | Alireza Taravat |
collection | DOAJ |
description | Accurate spatial information of agricultural fields is important for providing actionable information to farmers, managers, and policymakers. On the other hand, the automated detection of field boundaries is a challenging task due to their small size, irregular shape and the use of mixed-cropping systems making field boundaries vaguely defined. In this paper, we propose a strategy for field boundary detection based on the fully convolutional network architecture called ResU-Net. The benefits of this model are two-fold: first, residual units ease training of deep networks. Second, rich skip connections within the network could facilitate information propagation, allowing us to design networks with fewer parameters but better performance in comparison with the traditional U-Net model. An extensive experimental analysis is performed over the whole of Denmark using Sentinel-2 images and comparing several U-Net and ResU-Net field boundary detection algorithms. The presented results show that the ResU-Net model has a better performance with an average F<sub>1</sub> score of 0.90 and average Jaccard coefficient of 0.80 in comparison to the U-Net model with an average F<sub>1</sub> score of 0.88 and an average Jaccard coefficient of 0.77. |
first_indexed | 2024-03-09T00:49:33Z |
format | Article |
id | doaj.art-6aca3217d22841078e0546f5326b7c92 |
institution | Directory Open Access Journal |
issn | 2072-4292 |
language | English |
last_indexed | 2024-03-09T00:49:33Z |
publishDate | 2021-02-01 |
publisher | MDPI AG |
record_format | Article |
series | Remote Sensing |
spelling | doaj.art-6aca3217d22841078e0546f5326b7c922023-12-11T17:17:40ZengMDPI AGRemote Sensing2072-42922021-02-0113472210.3390/rs13040722Advanced Fully Convolutional Networks for Agricultural Field Boundary DetectionAlireza Taravat0Matthias P. Wagner1Rogerio Bonifacio2David Petit3Deimos Space, Oxford OX11 0QR, UKPanopterra, 64293 Darmstadt, GermanyUnited Nations World Food Programme UN-WFP, 00148 Rome, ItalyDeimos Space, Oxford OX11 0QR, UKAccurate spatial information of agricultural fields is important for providing actionable information to farmers, managers, and policymakers. On the other hand, the automated detection of field boundaries is a challenging task due to their small size, irregular shape and the use of mixed-cropping systems making field boundaries vaguely defined. In this paper, we propose a strategy for field boundary detection based on the fully convolutional network architecture called ResU-Net. The benefits of this model are two-fold: first, residual units ease training of deep networks. Second, rich skip connections within the network could facilitate information propagation, allowing us to design networks with fewer parameters but better performance in comparison with the traditional U-Net model. An extensive experimental analysis is performed over the whole of Denmark using Sentinel-2 images and comparing several U-Net and ResU-Net field boundary detection algorithms. The presented results show that the ResU-Net model has a better performance with an average F<sub>1</sub> score of 0.90 and average Jaccard coefficient of 0.80 in comparison to the U-Net model with an average F<sub>1</sub> score of 0.88 and an average Jaccard coefficient of 0.77.https://www.mdpi.com/2072-4292/13/4/722deep learningfully convolutional neural networksimage segmentationfield boundary detectioncropland monitoringU-Net |
spellingShingle | Alireza Taravat Matthias P. Wagner Rogerio Bonifacio David Petit Advanced Fully Convolutional Networks for Agricultural Field Boundary Detection Remote Sensing deep learning fully convolutional neural networks image segmentation field boundary detection cropland monitoring U-Net |
title | Advanced Fully Convolutional Networks for Agricultural Field Boundary Detection |
title_full | Advanced Fully Convolutional Networks for Agricultural Field Boundary Detection |
title_fullStr | Advanced Fully Convolutional Networks for Agricultural Field Boundary Detection |
title_full_unstemmed | Advanced Fully Convolutional Networks for Agricultural Field Boundary Detection |
title_short | Advanced Fully Convolutional Networks for Agricultural Field Boundary Detection |
title_sort | advanced fully convolutional networks for agricultural field boundary detection |
topic | deep learning fully convolutional neural networks image segmentation field boundary detection cropland monitoring U-Net |
url | https://www.mdpi.com/2072-4292/13/4/722 |
work_keys_str_mv | AT alirezataravat advancedfullyconvolutionalnetworksforagriculturalfieldboundarydetection AT matthiaspwagner advancedfullyconvolutionalnetworksforagriculturalfieldboundarydetection AT rogeriobonifacio advancedfullyconvolutionalnetworksforagriculturalfieldboundarydetection AT davidpetit advancedfullyconvolutionalnetworksforagriculturalfieldboundarydetection |