Field Boundary data can change the future of British farming

Agrimetrics, one of four agri-tech centres at the heart of the UK Government's Agricultural Technologies strategy, have used artificial intelligence and satellite imagery 13 times the resolution of the industry standard to identify 2.8 million field boundaries across England, Wales, Scotland and Northern Ireland.

Professor Richard Tiffin, Agrimetrics Chief Scientific Officer, explains the reason behind this:

"In addition to vast archives of historical information, our sector is capturing huge volumes of new data every day. Unfortunately, this data is rarely filtered by field level. As a result, drawing insights or creating products which can improve land management is often not possible – despite the required data being available."

Field Boundaries act as a cookie-cutter for this data. Data sets which could previously only be viewed by region or county can now be viewed from the perspective of an individual field. This has profound implications for a range of stakeholders.

Software developers will be able to improve user experience, input manufacturers can increase the efficacy and sustainability of their products and researchers can undertake innovative and important research – producing practical insights which improve farm management.

Accessing Field Boundaries through Agrimetrics brings several additional benefits, as Tiffin explains further:

"Field Boundaries provide a detailed digital map of the UK's farmed landscape, what's less well known is that they also provide a framework for organising and assembling data and the foundations for building new products."

"Here, Agrimetrics have provided a short-cut. We have linked our 2.8 million field boundaries to over a billion additional data points, including weather, previous cropping and soil composition. This expands the applications of field boundaries and provides significant time- and cost-saving for our users."

"When the required information is not available through Agrimetrics, Field Boundaries provide the building blocks for assembling and analysing third party data. Processing satellite data is one example. Field Boundaries act as cookie cutters to slice up satellite imagery in order to provide summary statistics for all of the individual fields in the UK."

"This will enable the development of products which could measure in-season crop health and productivity - without ever having to visit a field."

The second key differentiator of Agrimetrics Field Boundaries is their level of precision, explained here by Kathryn Berger, who leads the data science team responsible for bringing Field Boundaries to life:

"To create field boundaries, we trained a machine-learning algorithm to look at satellite data, identify the land features which distinguish fields, and use these features to highlight the field boundaries."

"Where Agrimetrics differs from other providers, however, is in the precision of the satellite data we use. Whereas industry-standard satellite imagery might have a spatial resolution of 10 to 20 metres, we used premium SPOT satellite imagery supplied by Airbus, which has a resolution of just 1.5 metres."

"The increased precision of our source data gives our algorithm a distinct advantage when identifying field boundaries."