Delineation of agricultural fields in smallholder farms from satellite images using fully convolutional networks and combinatorial grouping

Research output: Contribution to journalArticleAcademicpeer-review

1 Citation (Scopus)
36 Downloads (Pure)

Abstract

Accurate spatial information of agricultural fields in smallholder farms is important for providing actionable information to farmers, managers, and policymakers. Very High Resolution (VHR) satellite images can capture such information. However, the automated delineation of fields in smallholder farms is a challenging task because of their small size, irregular shape and the use of mixed-cropping systems, which make their boundaries vaguely defined. Physical edges between smallholder fields are often indistinct in satellite imagery and contours need to be identified by considering the transition of the complex textural pattern between fields. In these circumstances, standard edge-detection algorithms fail to extract accurate boundaries. This article introduces a strategy to detect field boundaries using a fully convolutional network in combination with a globalisation and grouping algorithm. The convolutional network using an encoder-decoder structure is capable of learning complex spatial-contextual features from the image and accurately detects sparse field contours. A hierarchical segmentation is derived from the contours using the oriented watershed transform and by iteratively merging adjacent regions based on the average strength of their common boundary. Finally, field segments are obtained by adopting a combinatorial grouping algorithm exploiting the information of the segmentation hierarchy. An extensive experimental analysis is performed in two study areas in Nigeria and Mali using WorldView-2/3 images and comparing several state-of-the-art contour detection algorithms. The algorithms are compared based on the precision-recall accuracy assessment strategy which is tolerating small localisation errors in the detected contours. The proposed strategy shows promising results by automatically delineating field boundaries with F-scores higher than 0.7 and 0.6 on our two test areas, respectively, outperforming alternative techniques.

Original languageEnglish
Article number111253
Pages (from-to)1-18
Number of pages18
JournalRemote sensing of environment
Volume231
Early online date18 Jun 2019
DOIs
Publication statusPublished - 15 Sep 2019

Fingerprint

small farms
smallholder
Farms
Satellites
farm
field margin
segmentation
Satellite imagery
accuracy assessment
Edge detection
Watersheds
Merging
satellite imagery
globalization
cropping practice
mixed cropping
Managers
transform
learning
satellite image

Keywords

  • Convolutional neural networks
  • Deep learning
  • Field boundary detection
  • Image segmentation
  • Semantic edge detection
  • Smallholder farming
  • ITC-ISI-JOURNAL-ARTICLE
  • ITC-HYBRID

Cite this

@article{2a6095b189a34767b2887a2e27f75f48,
title = "Delineation of agricultural fields in smallholder farms from satellite images using fully convolutional networks and combinatorial grouping",
abstract = "Accurate spatial information of agricultural fields in smallholder farms is important for providing actionable information to farmers, managers, and policymakers. Very High Resolution (VHR) satellite images can capture such information. However, the automated delineation of fields in smallholder farms is a challenging task because of their small size, irregular shape and the use of mixed-cropping systems, which make their boundaries vaguely defined. Physical edges between smallholder fields are often indistinct in satellite imagery and contours need to be identified by considering the transition of the complex textural pattern between fields. In these circumstances, standard edge-detection algorithms fail to extract accurate boundaries. This article introduces a strategy to detect field boundaries using a fully convolutional network in combination with a globalisation and grouping algorithm. The convolutional network using an encoder-decoder structure is capable of learning complex spatial-contextual features from the image and accurately detects sparse field contours. A hierarchical segmentation is derived from the contours using the oriented watershed transform and by iteratively merging adjacent regions based on the average strength of their common boundary. Finally, field segments are obtained by adopting a combinatorial grouping algorithm exploiting the information of the segmentation hierarchy. An extensive experimental analysis is performed in two study areas in Nigeria and Mali using WorldView-2/3 images and comparing several state-of-the-art contour detection algorithms. The algorithms are compared based on the precision-recall accuracy assessment strategy which is tolerating small localisation errors in the detected contours. The proposed strategy shows promising results by automatically delineating field boundaries with F-scores higher than 0.7 and 0.6 on our two test areas, respectively, outperforming alternative techniques.",
keywords = "Convolutional neural networks, Deep learning, Field boundary detection, Image segmentation, Semantic edge detection, Smallholder farming, ITC-ISI-JOURNAL-ARTICLE, ITC-HYBRID",
author = "C. Persello and Tolpekin, {V. A.} and Bergado, {J. R.} and {de By}, {R. A.}",
year = "2019",
month = "9",
day = "15",
doi = "10.1016/j.rse.2019.111253",
language = "English",
volume = "231",
pages = "1--18",
journal = "Remote sensing of environment",
issn = "0034-4257",
publisher = "Elsevier",

}

Delineation of agricultural fields in smallholder farms from satellite images using fully convolutional networks and combinatorial grouping. / Persello, C.; Tolpekin, V. A.; Bergado, J. R.; de By, R. A.

In: Remote sensing of environment, Vol. 231, 111253, 15.09.2019, p. 1-18.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Delineation of agricultural fields in smallholder farms from satellite images using fully convolutional networks and combinatorial grouping

AU - Persello, C.

AU - Tolpekin, V. A.

AU - Bergado, J. R.

AU - de By, R. A.

PY - 2019/9/15

Y1 - 2019/9/15

N2 - Accurate spatial information of agricultural fields in smallholder farms is important for providing actionable information to farmers, managers, and policymakers. Very High Resolution (VHR) satellite images can capture such information. However, the automated delineation of fields in smallholder farms is a challenging task because of their small size, irregular shape and the use of mixed-cropping systems, which make their boundaries vaguely defined. Physical edges between smallholder fields are often indistinct in satellite imagery and contours need to be identified by considering the transition of the complex textural pattern between fields. In these circumstances, standard edge-detection algorithms fail to extract accurate boundaries. This article introduces a strategy to detect field boundaries using a fully convolutional network in combination with a globalisation and grouping algorithm. The convolutional network using an encoder-decoder structure is capable of learning complex spatial-contextual features from the image and accurately detects sparse field contours. A hierarchical segmentation is derived from the contours using the oriented watershed transform and by iteratively merging adjacent regions based on the average strength of their common boundary. Finally, field segments are obtained by adopting a combinatorial grouping algorithm exploiting the information of the segmentation hierarchy. An extensive experimental analysis is performed in two study areas in Nigeria and Mali using WorldView-2/3 images and comparing several state-of-the-art contour detection algorithms. The algorithms are compared based on the precision-recall accuracy assessment strategy which is tolerating small localisation errors in the detected contours. The proposed strategy shows promising results by automatically delineating field boundaries with F-scores higher than 0.7 and 0.6 on our two test areas, respectively, outperforming alternative techniques.

AB - Accurate spatial information of agricultural fields in smallholder farms is important for providing actionable information to farmers, managers, and policymakers. Very High Resolution (VHR) satellite images can capture such information. However, the automated delineation of fields in smallholder farms is a challenging task because of their small size, irregular shape and the use of mixed-cropping systems, which make their boundaries vaguely defined. Physical edges between smallholder fields are often indistinct in satellite imagery and contours need to be identified by considering the transition of the complex textural pattern between fields. In these circumstances, standard edge-detection algorithms fail to extract accurate boundaries. This article introduces a strategy to detect field boundaries using a fully convolutional network in combination with a globalisation and grouping algorithm. The convolutional network using an encoder-decoder structure is capable of learning complex spatial-contextual features from the image and accurately detects sparse field contours. A hierarchical segmentation is derived from the contours using the oriented watershed transform and by iteratively merging adjacent regions based on the average strength of their common boundary. Finally, field segments are obtained by adopting a combinatorial grouping algorithm exploiting the information of the segmentation hierarchy. An extensive experimental analysis is performed in two study areas in Nigeria and Mali using WorldView-2/3 images and comparing several state-of-the-art contour detection algorithms. The algorithms are compared based on the precision-recall accuracy assessment strategy which is tolerating small localisation errors in the detected contours. The proposed strategy shows promising results by automatically delineating field boundaries with F-scores higher than 0.7 and 0.6 on our two test areas, respectively, outperforming alternative techniques.

KW - Convolutional neural networks

KW - Deep learning

KW - Field boundary detection

KW - Image segmentation

KW - Semantic edge detection

KW - Smallholder farming

KW - ITC-ISI-JOURNAL-ARTICLE

KW - ITC-HYBRID

UR - https://ezproxy2.utwente.nl/login?url=https://library.itc.utwente.nl/login/2019/isi/persello_del.pdf

U2 - 10.1016/j.rse.2019.111253

DO - 10.1016/j.rse.2019.111253

M3 - Article

VL - 231

SP - 1

EP - 18

JO - Remote sensing of environment

JF - Remote sensing of environment

SN - 0034-4257

M1 - 111253

ER -