File(s) not publicly available
Reason: Commercial licensing of imagery data and confidentiality of labelled data
Underlying data for 'Replacing Human Interpretation of Agricultural Land in Afghanistan with a Deep Convolutional Neural Network'
online resourceposted on 04.01.2021, 14:36 by Alex HamerAlex Hamer, Daniel SimmsDaniel Simms, Toby WaineToby Waine
The following data were used in the scientific paper 'Replacing Human Interpretation of Agricultural Land in Afghanistan with a Deep Convolutional Neural Network', which aims to determine whether CNNs can perform the role of a human interpreter in delineating agricultural land from medium-resolution imagery.
The DMC imagery (32m spatial resolution) used in this study are listed below with corresponding image dates and can be accessed from http://www.dmcii.com/.
- 27 April 2007 (du000aa3t)
- 26 March 2008 (dn000cac)
- 7 April 2008 (dn0005c9)
- 24 April 2008 (du000ccb)
- 25 March 2009 (du000e9at)
- 3 April 2009 (dn00078et)
- 8 April 2009 (dn000793t)
The agricultural masks used as the labelled dataset in this study were previously created from Taylor et al. (2010).
Taylor, J.C., Waine, T.W., Juniper, G.R., Simms, D.M. & Brewer, T.R. Survey and monitoring of opium poppy and wheat in Afghanistan: 2003-2009. Remote Sensing Letters, 1 (3), 179-185. doi:10.1080/01431161003713028