Projects


AI-based Cloud Mask Processor for 

Sentinel-2

Develop new and better cloud mask than existing ones

Duration: 2020–2021
Funder: European Space Agency

Planned outcomes:

  • Reliable cloud mask processor for Northern Europe region, which is compatible with ESA Sentinel-2 L2 processing chain.
  • Create high quality reference dataset for future developments.
  • Use innovative deep learning techniques in cloud masking.
 
 

Grazing detection from Copernicus data for agricultural subsidy checks

A complete satellite-based grassland monitoring service

Duration: 2020–2021
Funder: European Space Agency
Partners: Gisat s.r.o

Planned outcomes:

  • Develop grazing detection methodology based on Copernicus data (Sentinel-1 and Sentinel-2 imagery time series).
  • Provide the NPA operators means to carry out checks on grasslands using EO data and substituting on the spot checks of grassland grazing activity with new EO data based grazing detection methodology.
  • Close the grasslands subsidy checks case for CAP satellite monitoring.
Grazing detection project website
 
 

Harvesting Time Recommendation for maximum crop Yield (HaTRY)

Predicting the best time to harvest crop using remotely sensed data

Duration: 2020–2021
Funder: European Space Agency
Partners: local farmers in Estonia

Planned outcomes:

  • Prototype service for Precision Farming application to predict and recommend the most favorable harvesting time for maximum crop yield.
  • Focus on Northern Europe region and three most common crops: winter wheat, spring barley and winter rapeseed.
  • Learn and collect farmers requirements for a fully operational service.
 
 

National Programme for Addressing Socio-Economic Challenges through R&D (RITA). Using remote sensing data in favour of the public sector services

Monitoring the use of agricultural land

Duration: 2019–2020
Funder: This study was financially supported by the European Regional Development Fund within National Programme for Addressing Socio-Economic Challenges through R&D (RITA).
Partners: Consortium team: University of Tartu, Tallinn University of Technology, Estonian University of Life Sciences, KappaZeta Ltd

Planned outcomes:

  • Mature, reliable and tested crop classification methodology development specifically suited for Estonian agricultural, ecological and climatic conditions.
  • Multi-year country-wide testing and error analysis about vegetative seasons 2018–2020.
  • Crop classification model prototype with test datasets.

 

ESA Business Incubation

Improving our software quality during the incubation phase

Duration: 2018–2019 

Funder: European Space Agency
Partners: University of Tartu (Institute of Computer Science), Sookolli Kaardid OÜ, Elmer SKB OÜ, Codeborne OÜ
Outcomes:

  • Satellite imagery model development
  • Software architecture definition and review
  • Software implementation
  • Market analysis and business plan development
  • Improved web map for visualizing our analysis results
 

Detection of mowing events on grasslands from Sentinel-data

The first nation-wide system for automated monitoring of agricultural practices in EU 

Duration: 2016-2018
Customer: Estonian Agricultural Registers and Information Board (ARIB)
Partners: CGI Estonia, Tartu Observatory
Outcomes:

  • Nation-wide fully automated mowing detection system operational in Estonia from 2018
  • Sentinel-1 and Sentinel-2 time-series for operational near real-time monitoring
  • 85% of detection accuracy of the mowing events on grasslands
  • Automated “early warning” reminders to applicants



 An isolated grassland – status “not mowed”


Grassland mowing detection for agricultural subsidy checks with Sentinel-1 and Sentinel-2

Looking over the borders to Denmark, Sweden and Poland 

Duration: 2017-2019.
Customer: European Space Agency, Industry Incentive Scheme (ESA IIS)
Partners: The Danish Agrifish Agency, the Swedish Board of Agriculture, The Agency for Restructuring and Modernisation of Agriculture in Poland, Reach-U Ltd, KappaZeta Ltd
Outcomes:

  • Validate the service by performing user trials in Sweden and Denmark
  • Enhance the existing cutting and grazing detection methodology
  • Study Earth Observation Community Platform (EO CP) service providers
  • Perform Viability Analysis





Data Analytics for Optimizing Agricultural Monitoring

Going beyond the mowing detection powered by high-level data analytics expertise 

Duration: 2017-2018.
Funder: Enterprise Estonia, project No. EU48684
Partners: Software Technology and Applications Competence Centre (STACC), KappaZeta
Outcomes:

  • Develop a scientifically validated methodology for ploughing and grazing events detection from Sentinel-1 and Sentinel-2 time series
  • Bring to customers a new cultivation and grazing detection product prototype
  • Involve high-level data analytics expertise powered by STACC.



Home page and demo application

KZ branding, home page and demo application

Duration: 2017
Funder: Enterprise Estonia, project No. EU51738; amount of the grant 14,960 €.
Main partners: Reach-U, Kiften, Jon & Pun, KappaZeta
Outcomes: KZ branding, home page and mowing detection demo application



Research

Relating Sentinel-1 Interferometric Coherence to Mowing Events on Grasslands

  • the effect of grassland mowing on interferometric coherence calculated for a Sentinel-1 pair separated by 12 days
  • The coherence increases after a mowing event, but the effect wears off in a couple of weeks
  • Precipitation counteracts this effect

Go to paper (MDPI)



Monitoring of Agricultural Grasslands With Time Series of X-Band Repeat-Pass Interferometric SAR

  • Using COSMO-SkyMed acquisition pairs separated by one day, the relationship between grassland mowing and interferometric coherence was noticed
  • Precipitation and further agricultural activity on parcels after mowing may negate the coherence increase

Go to paper (IEEE)


Observations of Cutting Practices in Agricultural Grasslands Using Polarimetric SAR

  • SAR polarimetry also provides means to detect mowing events from TerraSAR-X data products
  • Polarimetric parameters such as HH/VV polarimetric coherence and alpha angle of the H/alpha decomposition are sensitive to mowing

Go to paper (MDPI)




 

Sensitivity of Sentinel-1 backscatter to
characteristics of buildings

  • Sentinel-1 backscatter statistics respect to the physical parameters of the buildings
  • Backscatter dependence on building height, material, orientation angle and shape
  • Alignment effects due to Sentinel-1 orbit and look angle

Go to paper (Taylor & Francis)