CropHype Improving crop recognition based on hyperspectral EnMAP data using the Geo Engine [funded by DLR]

Project staff:


Sebastian Egli
Leander Leist

Abstract:

Die Wirtschaft Kenias ist erheblich durch die Agrarproduktion geprägt. Kleinbäuerliche Strukturen sind dabei zentral für die lokale Ernährungssicherheit (Photo: S. Egli).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Kenya's economy is significantly shaped by agricultural production. Smallholder structures are central to local food security (Photo: S. Egli).

  

Smallholder farms are a central pillar of local food security in sub-Saharan Africa. Climate and environmental change are increasingly threatening the yields and thus the livelihoods of smallholders. An important element in reducing poverty and adapting to climate change is therefore to sustainably improve and secure the yield situation.

In order to develop appropriate measures, however, there is still a lack of basic data on crops and yields and how they change over time. One reason is the spectral and spatial resolution of the current operational environmental satellites, which is poorly adapted to smallholder structures.

The project "CropHype - Improving crop identification based on hyperspectral EnMAP data using the Geo Engine" therefore aims to make intensive use of the new hyperspectral data from the German EnMAP mission. From the fusion of multi- (Sentinel-2) and hyperspectral (EnMAP) data, a significantly improved classification of smallholder crops for different phenological phases is to be developed with the help of machine learning methods.

Important here is the cooperation with Geo Engine GmbH, via whose cloud service the technical implementation of the data connection, data pre-processing and data analysis will be implemented.

The associated partner agriBORA GmbH guarantees the connection to the smallholders on site. The field recordings of smallholder cultivation by agriBORA provide unique data for the optimal training of the classification algorithms. The results of the classification in turn help agriBORA to improve its crop forecasts and thus to be able to provide smallholders with optimal advice.

Translated with www.DeepL.com/Translator (free version)

 






Quick search

  • Publications:
  • Datasets: