Talk to us

Rural Payments Agency (RPA) customer story

Rural Payments Agency (RPA) leverages Picterra’s GeoAI to enhance sustainable farming initiatives

The Rural Payments Agency (RPA), an executive agency of the UK Department for Environment, Food and Rural Affairs (DEFRA), plays a pivotal role in supporting sustainable farming practices and rural economic growth across the UK. Tasked with administering up to £2 billion annually in agricultural and rural development payments, RPA ensures the delivery of schemes such as Countryside Stewardship (CS) and Sustainable Farming Incentive (SFI). By distributing funds and providing guidance to farmers, landowners, and rural businesses, RPA aims to boost farming productivity, ensure food provenance, control livestock diseases, and support environmental outcomes.

This customer story explores how RPA has begun to transform how it monitors schemes, delivers its regulatory requirements, and supports habitat monitoring with Picterra’s platform. It highlights the challenges the Agency faced, the solutions implemented, and the remarkable results achieved.

Solution:

National habitat monitoring for biodiversity and sustainability

Industry:

Agriculture and environmental management

Screenshot of the model training process within Picterra showing training (yellow), testing (red), and accuracy (blue) areas and their corresponding annotations and results.

We needed to become more self-reliant in creating our own datasets. The RPA administers the farming schemes and maintains the Rural Land Register (RLR), a digital land dataset covering agricultural parcel boundaries and their land cover habitats. RLR is maintained to a 3-year currency for any potential change to ensure we make accurate payments to our customers.

Yajnaseni Palchoudhuri

Earth Observation Specialist, Rural Payment Agency DEFRA

Challenge

The UK’s transition from the European Union’s Common Agricultural Policy (CAP) to the Environmental Land Management schemes (ELM) meant that RPA faced the challenge of rethinking how it monitors agreements to meet new environmental priorities and regulations. The revised 25-Year Environment Plan (YEP), introduced in 2018, emphasized sustainability and net-zero commitments, making the need for innovative monitoring solutions more relevant.

The Rural Payments Agency (RPA) operates in a regulated environment where ensuring compliance with agricultural and environmental standards is crucial. With the new Environment Land Management schemes, RPA is transforming into a data-led organization, harnessing the power of intelligence to help customers get the most from their agreements and support robust decision-making. In terms of monitoring, this is taking the Agency from small-scale sampling and imaging for scheme controls to a fully integrated, national, and continuous large-area proactive monitoring system. Large-scale satellite analytics in terms of machine learning and AI is required to provide continuous monitoring of the landscape within the farm across the season.

Before partnering with Picterra, RPA faced several significant challenges:

Reliance on manual inspection and limited coverage:
Reliance on manual inspection and limited coverage:
  • Traditional methods for monitoring farming practices were resource-intensive and time-consuming. RPA relied on random, small-scale sampling and verification in the field, resulting in limited or inconsistent coverage.

  • The extensive manual effort required for data collection and analysis through field inspections meant that only a fraction of the land could be assessed yearly.
Image description
Ensuring compliance with regulations
Ensuring compliance with regulations
  • Farmers and land managers must adhere to the specific prescriptions of options and actions included within CS and SFI, such as controlling the use of fertilizers, creating buffer strips around fields, and maintaining valuable areas for biodiversity.

  • For delivering successful outcomes to the new agri-environment schemes, RPA intends to implement a checks-by-monitoring approach across the season, applied to the whole population of agreements for early identification and prevention of errors and to support farmers to get the most from their agreements. The large-scale area monitoring was challenging using just field visits, resulting in an increased number of scheme irregularities. The use of technology will ensure robust, scalable monitoring across the country, thereby reducing such irregularities efficiently.
Image description
Environmental and biodiversity goals
Environmental and biodiversity goals
  • RPA needed to support and promote sustainable farming practices that align with environmental conservation, climate action, and biodiversity enhancement. Hedgerows, being a critical component of the UK’s rural landscape, are included within the new Hedge management schemes as well as Hedge regulations, which require the land managers to develop and maintain the hedges as per the official guidance.

  • To monitor the management activities on the rural Hedges across England, the agency required a baseline national hedge and tree map to cover the presence and the dimensions of hedges and their seasonal changes across the country. They needed an automated machine learning tool to effectively map and monitor these farm habitats and other key indicators of a balanced ecosystem, such as individual trees and buffer strips.
Image description
Technological limitations
Technological limitations
  • Existing tools and workflows were inefficient for the high-resolution land parcel habitat monitoring needed to meet RPA’s goals.

  • The existing data sources were inadequate and inconsistent for a dynamic regulatory framework.
Image description

To address these challenges, RPA sought a solution that could automate and enhance its habitat monitoring capabilities, provide accurate and timely data, and support its broader goals of promoting sustainable farming and biodiversity enhancement. This need for innovation and efficiency set the stage for its partnership with Picterra.

The models developed using Picterra's platform achieved an accuracy of 85-90%, consistent across different regions. This consistency is reassuring and supports the platform's reliability.

Yajnaseni Palchoudhuri

Earth Observation Specialist, Rural Payment Agency DEFRA

Solution

The collaboration between RPA and Picterra provided a comprehensive solution to enhance RPA’s monitoring capabilities and streamline their workflows. 

Here’s how Picterra’s platform addressed RPA’s needs:

Advanced geospatial analytical framework including machine learning
  • Picterra’s platform leverages state-of-the-art AI and deep learning technologies to analyze high-resolution aerial and satellite imagery. This enables accurate detection and mapping of key features such as hedgerows, individual trees, bare soil, and nesting plots.

  • The platform’s deep learning algorithms efficiently create large-scale maps to locate habitat features, providing RPA with precise and up-to-date information critical for the FCP scheme and regulatory compliance monitoring.
Improved hedgerow mapping
  • One primary focus of the collaboration was to enhance hedgerow mapping across the country. Picterra’s platform detected and mapped hedgerows and trees with an 85-90% accuracy rate.

  • The hedgerow and tree map is built using aerial imagery from the Department for Science, Innovation and Technology (DSIT)’s Aerial Photography of Great Britain (APGB) program, which operates on a two-year cycle to cover all of England. RPA plans to update this product on the same two-year cycle to maintain an up-to-date Hedge Control dataset. High-frequency satellite imagery may be used as a reference for cross-validation during this process.
Integration with existing workflows
  • Picterra’s platform seamlessly integrates with RPA’s existing tools and platforms, including QGIS, Safe FME, and ArcGIS. This ensured compatibility and allowed RPA to enhance their workflows without significant disruptions.

  • The platform’s robust API capabilities facilitated smooth data exchange and integration, enabling RPA to incorporate Picterra’s outputs into their broader automated workflows.
Scalable and efficient monitoring
  • Picterra’s solution provided better scalability and efficiency by moving processes to the cloud. RPA could now handle larger volumes of data and perform extensive analysis without the limitations of on-premise infrastructure.

  • The collaboration involved training the model with thousands of labeled examples of hedges and trees from rural and urban areas, enabling the detector to identify hedgerows and trees across England. Production is underway to create a baseline national tree and hedge layer for 2024. Once this layer is delivered, it will be used to maintain the RPA Hedge Control layer as well as to validate the Hedge scheme requirements. The baseline national map will also be used across DEFRA to support the assessment of terrestrial natural capital assets.
Improved data accuracy and frequency
  • Picterra’s solution can integrate various high-resolution data sources, including Skysat, Planet Fusion, APGB, Worldview, Quickbird, and airborne multispectral imagery. This diverse data input ensured comprehensive and enhanced mapping accuracy.
Support for sustainable farming initiatives
  • Picterra’s platform supported the goals of the Sustainable Farming Incentive (SFI) and other initiatives by providing detailed and accurate mapped data on farm habitats. This mapped habitat data will support RPA to monitor sustainable farming practices in line with environmental land management agreements.

  • The advanced AI models enabled RPA to automate the feature recognition process to identify their presence and monitor and manage key indicators of a balanced ecosystem, such as hedgerows, trees, and buffer zones.
Screenshot of a report displaying the results of running a trained model on satellite imagery.

Picterra’s platform has revolutionized our approach to monitoring and managing agricultural practices. The remarkable accuracy and efficiency improvements have allowed us to achieve our environmental and compliance goals more effectively.

Yajnaseni Palchoudhuri

Earth Observation Specialist, Rural Payment Agency DEFRA

Conclusion

The partnership between RPA and Picterra has transformed how RPA monitors agricultural practices and ensures compliance with environmental schemes. By leveraging Picterra’s advanced geospatial machine learning platform, RPA has improved monitoring accuracy, increased operational efficiency, and strengthened its support for sustainable farming.

The collaboration has received strong positive feedback, and RPA plans to expand the use of Picterra’s platform for future monitoring initiatives.

Through this work, RPA has set a new benchmark for agricultural monitoring, demonstrating how geospatial technologies can play a vital role in delivering effective, scalable, and sustainable land management.

Results

85-90% accuracy

in mapping hedgerows, trees, and buffer zones

Increased efficiency

through AI-driven automation and scaling nationwide

Support for sustainability

mapping biodiversity and environmental goals

Power your sustainability ecosystem with GeoAI

Talk to our experts

Source responsibly

Monitor compliance

Detect risks

Surface intelligence

Picterra Forge

Picterra Tracer

Picterra Accelerate

Integrations

Impact

Partnership

Resource Hub

Blog

Documentation

Talk to us