Identifying crops in smallholder farms using time series of WorldView-2 images

R. Zurita-Milla, E. Izquierdo-Verdiguier, R.A. de By

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

1 Citation (Scopus)

Abstract

A high heterogeneity in farming factors (soils, weather, inputs, practice) characterizes the typical smallholder farm landscapes of sub-Saharan Africa. This complicates automatic classification to crop when using only spectral information of very high spatial resolution image time series. This work addresses the crop identification problem in smallholder landscapes through three steps: features extraction, feature selection and classification. Feature extraction is used to exted the spatial-spectral information of the farm fields, with a substantial number of features considered through cloud computing. Feature selection is based on correlation between the features and the labels of the field's crops and it is applied to reduce the dimensionality of the data without lose information. Finally, a random forest classifier is applied to identify a crop class per field. Good preliminary results were obtained reducing the number of features from 1638 to 66. The overall accuracy achieves 80% in the test set using a random forest classifier.
Original languageEnglish
Title of host publication9th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2017)
Place of PublicationPiscataway, NJ
PublisherIEEE
Number of pages3
ISBN (Electronic)978-1-5386-3327-4
ISBN (Print)978-1-5386-3328-1
DOIs
Publication statusPublished - 2017
Event9th International Workshop on the Analysis of Multi-temporal Remote Sensing Images, MultiTemp 2017 - Bruges, Belgium
Duration: 26 Jun 201729 Jun 2017

Conference

Conference9th International Workshop on the Analysis of Multi-temporal Remote Sensing Images, MultiTemp 2017
Abbreviated titleMultiTemp
CountryBelgium
CityBruges
Period26/06/1729/06/17

Fingerprint

smallholder
farm
time series
crop
spatial resolution
weather
WorldView
soil

Cite this

Zurita-Milla, R., Izquierdo-Verdiguier, E., & de By, R. A. (2017). Identifying crops in smallholder farms using time series of WorldView-2 images. In 9th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2017) Piscataway, NJ: IEEE. https://doi.org/10.1109/Multi-Temp.2017.8035246
Zurita-Milla, R. ; Izquierdo-Verdiguier, E. ; de By, R.A. / Identifying crops in smallholder farms using time series of WorldView-2 images. 9th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2017). Piscataway, NJ : IEEE, 2017.
@inbook{b9cf36b064f04856a1d6ef74dfe090ae,
title = "Identifying crops in smallholder farms using time series of WorldView-2 images",
abstract = "A high heterogeneity in farming factors (soils, weather, inputs, practice) characterizes the typical smallholder farm landscapes of sub-Saharan Africa. This complicates automatic classification to crop when using only spectral information of very high spatial resolution image time series. This work addresses the crop identification problem in smallholder landscapes through three steps: features extraction, feature selection and classification. Feature extraction is used to exted the spatial-spectral information of the farm fields, with a substantial number of features considered through cloud computing. Feature selection is based on correlation between the features and the labels of the field's crops and it is applied to reduce the dimensionality of the data without lose information. Finally, a random forest classifier is applied to identify a crop class per field. Good preliminary results were obtained reducing the number of features from 1638 to 66. The overall accuracy achieves 80{\%} in the test set using a random forest classifier.",
author = "R. Zurita-Milla and E. Izquierdo-Verdiguier and {de By}, R.A.",
year = "2017",
doi = "10.1109/Multi-Temp.2017.8035246",
language = "English",
isbn = "978-1-5386-3328-1",
booktitle = "9th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2017)",
publisher = "IEEE",
address = "United States",

}

Zurita-Milla, R, Izquierdo-Verdiguier, E & de By, RA 2017, Identifying crops in smallholder farms using time series of WorldView-2 images. in 9th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2017). IEEE, Piscataway, NJ, 9th International Workshop on the Analysis of Multi-temporal Remote Sensing Images, MultiTemp 2017, Bruges, Belgium, 26/06/17. https://doi.org/10.1109/Multi-Temp.2017.8035246

Identifying crops in smallholder farms using time series of WorldView-2 images. / Zurita-Milla, R.; Izquierdo-Verdiguier, E.; de By, R.A.

9th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2017). Piscataway, NJ : IEEE, 2017.

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

TY - CHAP

T1 - Identifying crops in smallholder farms using time series of WorldView-2 images

AU - Zurita-Milla, R.

AU - Izquierdo-Verdiguier, E.

AU - de By, R.A.

PY - 2017

Y1 - 2017

N2 - A high heterogeneity in farming factors (soils, weather, inputs, practice) characterizes the typical smallholder farm landscapes of sub-Saharan Africa. This complicates automatic classification to crop when using only spectral information of very high spatial resolution image time series. This work addresses the crop identification problem in smallholder landscapes through three steps: features extraction, feature selection and classification. Feature extraction is used to exted the spatial-spectral information of the farm fields, with a substantial number of features considered through cloud computing. Feature selection is based on correlation between the features and the labels of the field's crops and it is applied to reduce the dimensionality of the data without lose information. Finally, a random forest classifier is applied to identify a crop class per field. Good preliminary results were obtained reducing the number of features from 1638 to 66. The overall accuracy achieves 80% in the test set using a random forest classifier.

AB - A high heterogeneity in farming factors (soils, weather, inputs, practice) characterizes the typical smallholder farm landscapes of sub-Saharan Africa. This complicates automatic classification to crop when using only spectral information of very high spatial resolution image time series. This work addresses the crop identification problem in smallholder landscapes through three steps: features extraction, feature selection and classification. Feature extraction is used to exted the spatial-spectral information of the farm fields, with a substantial number of features considered through cloud computing. Feature selection is based on correlation between the features and the labels of the field's crops and it is applied to reduce the dimensionality of the data without lose information. Finally, a random forest classifier is applied to identify a crop class per field. Good preliminary results were obtained reducing the number of features from 1638 to 66. The overall accuracy achieves 80% in the test set using a random forest classifier.

UR - http://ShowEdit https://ezproxy2.utwente.nl/login?url=https://doi.org/10.1109/Multi-Temp.2017.8035246

UR - http://ezproxy.utwente.nl:2048/login?url=https://webapps.itc.utwente.nl/library/2017/conf/zuritamilla_ide.pdf

UR - https://ezproxy2.utwente.nl/login?url=https://webapps.itc.utwente.nl/library/2017/chap/zuritamilla_ide.pdf

U2 - 10.1109/Multi-Temp.2017.8035246

DO - 10.1109/Multi-Temp.2017.8035246

M3 - Chapter

SN - 978-1-5386-3328-1

BT - 9th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2017)

PB - IEEE

CY - Piscataway, NJ

ER -

Zurita-Milla R, Izquierdo-Verdiguier E, de By RA. Identifying crops in smallholder farms using time series of WorldView-2 images. In 9th International Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2017). Piscataway, NJ: IEEE. 2017 https://doi.org/10.1109/Multi-Temp.2017.8035246