Agricultural SandboxNL

A Crop Parcel Level Database Using Sentinel-1 SAR and Google Earth Engine

Conference Paper (2021)
Author(s)

Vineet Kumar (TU Delft - Civil Engineering & Geosciences)

Manuel Huber (European Space Agency (ESA))

Maurice Shorachi (Student TU Delft)

Bjorn Rommen (European Space Agency (ESA))

Susan C. Steele-Dunne (TU Delft - Civil Engineering & Geosciences)

Research Group
Mathematical Geodesy and Positioning
DOI related publication
https://doi.org/10.1109/IGARSS47720.2021.9553571 Final published version
More Info
expand_more
Publication Year
2021
Language
English
Research Group
Mathematical Geodesy and Positioning
Article number
9553571
Pages (from-to)
6284-6287
ISBN (print)
978-1-6654-4762-1
ISBN (electronic)
978-1-6654-0369-6
Event
IGARSS 2021 (2021-07-11 - 2021-07-16), Virtual at Brussels, Belgium
Downloads counter
185

Abstract

The systematic high temporal coverage of Sentinel-1 Synthetic Aperture Radar (SAR) is ideal for agricultural monitoring. The availability of these data on cloud computing infrastructure eliminates the need for massive computing power to process imagery. However, their distribution as SAR imagery still limits their accessibility for non-expert users. In Agricultural SandboxNL, Copernicus Sentinel-1 imagery on the Google Earth Engine (GEE) was mined to produce a database of spatially-tagged, parcel-level backscatter for every agricultural parcel in the Netherlands from 2017 to 2019. The database includes descriptors from the publicly available Basisregistratie Gewaspercelen, allowing the user to query the database by crop type and administrative boundary for any region of interest within The Netherlands. Publication of this database reduces the burden of processing and extracting a large volume of Sentinel-1 SAR data for experts. More importantly, it provides easy access to the Sentinel-1 data for agriculture/agronomy experts with limited SAR processing experience. In addition, the GEE package Sen1byParcel developed for Agricultural SandboxNL is made publicly available so that Sentinel-1 imagery can be extracted for any user-provided shapefile.