DIGIFARM

Detecting the world’s highest accuracy field boundaries to power precision agriculture

Using deep neural network models and super-resolution Sentinel-2 EO-Data for detecting the world’s highest accuracy field boundaries

DigiFarm is a Norwegian based ag-tech startup established in 2019. DigiFarm’s core vision is to detect the world’s most accurate field boundaries and seeded acres to power precision agriculture, leveraging the latest advancements in Artificial Intelligence technology. This is achieved through developing deep neural network models for automatically detecting field boundaries through super-resolving Sentinel-2 satellite imagery (to 1 meter resolution). DigiFarm has successfully validated the model on 15+ million hectares of fields achieving detection accuracies of above 96%, 12-15% higher than existing boundary data (Cadastral, LPIS in EU and CLUs in US). 

The project included developing and model training deep neural network model for detection of entire-country sized regions including: Germany, Austria, Belgium and the United Kingdom.

The solution will benefit agricultural organizations to optimize operations, making better data-driven decisions, reducing seasonal uncertainty, minimising production costs, increasing crop-revenue and enabling key development in Carbon capture leading to reduction in Co2 footprints. DigiFarm delivers software as a service seamlessly integrated into clients (B2B/B2G) digital solutions (API). The pilot results will be implemented into corporate partners (KWS) commercial FMS-solutions after the pilot has been finalized.

ACHIEVEMENTS

DigiFarm managed successfully to delineate field boundaries and seeded acres across the AOIs including: Germany, Austria, Belgium and the United Kingdom. DigiFarm managed to achieve the targeted accuracy (IoU) of 0.94+ across all the regions. This resulted in achieving 10-12% higher accuracy than LPIS (Land Parcel Identification System) Cadastral data which was benchmarked against our field boundary delineation assessed across 200k hectares in England, Belgium and Austria. 

The outcome included following field boundaries delineated:

 

  • Belgium
    • Total hectares: 8,234,991
    • Total polygons delineated: 1,105,376
    • IoU accuracy: 0.95

       

  • Germany
    • Total hectares: 46,763,674
    • Total polygons delineated: 4,418,336
    • IoU accuracy: 0.96

       

  • Austria
    • Total hectares: 12,264,239
    • Total polygons: 731,323
    • IoU accuracy: 0.95

       

  • United Kingdom
    • Total hectares: 18,409,078
    • Total polygons: 2,502,201
    • IoU accuracy: 0.95

HOW THE EOSC SERVICES WERE USED

Digifarm used the EOSC services in order to train the deep neural network model to automatically and accurately detect agricultural field boundaries based on the deep-resolution Sentinel-2 at 1m per pixel resolution Satellite Earth Observation data. In order to be able to train and develop a highly accurate AI-model (image segmentation) it was critical for us to be able to leverage the HPC and high-performance GPUs in order to train the models, this would not have been possible to do in a commercial setting as it’s too expensive, hence, the ability to leverage this partnership along with the internal team’s expertise was critical to DigiFarm success.

THE VALUE PROPOSAL OF THE PILOT

All precision farming services and in-field analytics start with accurate field boundaries and seeded acres. Unfortunately, existing field boundary data is outdated and inaccurate. Large scale boundary data is managed through national agencies (Cadastral) such as the Land Parcel Identification System for all 27 EU-regions (36 million boundaries) and the Common Land Units in the US (32 million boundaries) which are over 14 years old and have not been updated since. Field boundaries are still updated by people manually digitizing field boundaries on a map, which is inaccurate, time consuming and costly.  

As all precision farming and in-field analytics, the solution we have developed benefits agricultural organizations to optimize operations, making better data-driven decisions, reducing seasonal uncertainty, minimizing production costs, increasing crop-revenue and enabling key development in Carbon capture leading to reduction in Co2 footprints. DigiFarm delivers software as a service seamlessly integrated into clients (B2B/B2G) digital solutions (API).

HOW EOSC helped

The EOSC team’s help was crucial for us to be able to get setup with the HPC-kit, including onboarding and providing technical assistance on how to leverage and optimize our model training and GPU-setup, this was important for us to be able to build and develop a scalable, automatic and cost-efficient data processing pipeline. EGI-ACE contributed with the provisioning of free at the point of use access to GPU-enabled cloud resources at INFN-CNAF.

 

 

 

FUTURE PLANS

Digifarme achieved its targets in the EOSC FUTURE project and have since the close of the project managed to secure additional commercial contracts with multiple corporate partners in the agricultural sector including: Bayer Crop Science, Corteva, Lithuania Agricultural Paying Agency, Limagrain, E-stratos, Agricolus, Klim, Agdir and Aganalyst.

Future plans will include expanding to cover all of the 27 EU-regions in addition to expanding to India and parts of Africa including Tanzania, Ethiopia in order to help smallholder farmers optimize their grain-production with local partners.

Business Partner

Service provider

Supporting project