Secretariat

Anja Herkner
Institute of Rural Studies

Bundesallee 64
38116 Braunschweig
Phone: +49 531 596 5501
Fax: +49 531 596 5599
lr@thuenen.de


Development of an automated support system to monitor agricultural funding sites (timeStamp)

Project

Identification of land use by Sentinel-1 satellite image time series (c) Bild links [image left]: www.esa.int -  Bild rechts [image right]: http://webapp.flf.julius-kuehn.de/gisapp/projekte
Recording of land-use by remote-sensing (© Bild links [image left]: www.esa.int - Bild rechts [image right]: http://webapp.flf.julius-kuehn.de/gisapp/projekte)

The monitoring of area-related measures in context of the Common Agricultural Policy (CAP) and nature conservation is currently carried out through time-consuming and cost-intensive on-site inspections. By using Sentinel satellite data, authorities of the agricultural and nature conservation agencies shall be able to reduce on-site inspections in the future. A Web application, based on remote sensing data, shall enable this. A cloud-based web application has developed for automated satellite data analysis and output of evaluation results to aid this process.

Background and Objective

The good spatial coverage and high temporal resolution of the Sentinel satellites of the Earth observation programme of Copernicus today allow a close temporal recording of agricultural land use and identification of land use changes.

We aim to aid administrative authorities and other interested users in utilisting satellite data to impove the efficiency of time- and cost-intensive on-site checks in the area of site monitoring by integrating satellite data anylsis into their workflows.

Within the framework of the project "timeStamp", scientists of the Thünen Institute and the other project partners are developing a web application. This application evaluates whether the declared land use corresponds with the land use depicted by satellite data, or if land use change is indicated.

A web application is developed for an exemplary use case " monitoring EFA/greening catch crop sites in the context of the EU agricultural funding (CAP)", it provides control cues with a traffic light colour coding related to the expected land use of catch cops.

In addition to this, a multifunctional basic function provides standard time-series graphs, test parameters, and change indicators usable for various areas of application related to land use change and management.

 

 

Target Group

Ministries of agriculture and environment, land authorities, official nature conservation, agricultural administration, European Commission

Approach

The following steps will be undertaken:

 

  • user workshops with cooperating authorities to assess the functionalities required and to determination the subject-specific and technical requirements of the web application to be developed.
  • Based on the requirements analysis, the processed satellite-image time series of Sentinel-2 are used to develop appropriate indicators and evaluation approaches (algorithms) and be validated by test data sets to monitor the lifetime of catch crops.
  • Development and graphics of indicators for the miltifunctional base function based on Sentinel-1 and -2-time series data.
  • The evaluation algorithms for the application are implenented into a cloud-based processing environment.
  • Implementation of an object-related evaluation of field and site geometries to be monitored based on remote sensing data.
  • Development of a user friendly web application to provide data exchange, selection of test indicators for the control sites and a spatial visualization of the results including site-specific information (e.g., a traffic-light system) for on-site controlls.
  • user workshop to inform potential users and get their feedback on usability and user experience

Data and Methods

By analyzing the Sentinel-1 (radar) and Sentinel-2 (spectral) time series data, algorithms for the identification of catch crops can be derived using appropriate metrics (measures) and indices for seeding, growth and plough of the catch crops.

For the development of the algorithms, catch-crop sites in the field are mapped and InVekoS and on-site control data are used as training and test data.

The technical infrastructure of the application is implemented in the Microsoft AzureCloud using docker containers as a modular backend. Via the user interface, users can create, send and manage test jobs as well as display and download the results.

Our Research Questions

  • What subject-specific requirements (required input data, metrics, indicators, temporal rhythms of the evaluations) have the respective authorities to the web application?
  • What are technical requirements for the target system (including data formats, software environments, interfaces, programming languages)?
  • What are appropriate metrics, indicators, and algorithms (e.g., NDVI, matching with phenological pattern profiles,) to recognize the lifetime of catch crops?

Preliminary Results

The prototype of the timeStamp application provides a cloud-based software infrastructure for automated remote analysis.

The user interface allows even users with only basic knowledge about remote sensing to integrate satellite data into their work.

The multifunctional basic function offers evaluation options for a variety of questions concerning land monitoring in environmental and nature conservation. The verification of catch crop sites is based on the analysis of the temporal pattern of the vegetation index (NDVI). From a set of test parameters (e.g. minima, maxima, trends) the probability of the presence of a catch crop is calculated. The site-specific evaluation results and the output of visual control instructions can support a targeted planning of spot checks.

Within the project 'Copernicus lights green' the application is to be developed further, migrated to CODE-DE and prepared for operational use.

Thünen-Contact


Involved Thünen-Partners


Involved external Thünen-Partners


Funding Body

  • Federal Ministry for Economic Affairs and Energy (BMWi)
    (national, öffentlich)
  • German Aerospace Center (DLR)
    (national, öffentlich)

Duration

8.2018 - 10.2021

More Information

Projekt type:
Project funding number: 50EW1704A
Funding program: DLR - Entwicklung und Implementierungsvorbereitung von Copernicus Diensten für den öffentlichen Bedarf in Deutschland
Project status: ongoing

Publications to the project

hits: 1

  1. Schulz C, Holtgrave A-K, Kleinschmit B (2021) Large-scale winter catch crop monitoring with Sentinel-2 time series and machine learning - An alternative to on-site controls? Comput Electron Agric 186:106173, DOI:10.1016/j.compag.2021.106173