IBM Research is working with British agricultural software specialist Farmplan and other partners to collect data from farmers to predict the spread of a damaging weed called Blackgrass. Participating farmers in the UK will provide data to train AI to predict just where the weed will grow.

Blackgrass costs farmers more than US $760 billion annually as it can compete with cereal crops such as wheat and barley reducing their yield and resulting in food shortages. As the weed is becoming increasingly resistant to herbicides scientists are turning to AI to fight it.

Working with farmers in the UK, a new £1 million project (aiScope) is in progress, with dozens of UK farmers and Farmplan working alongside IBM Research Europe, Sheffield University, Rothamsted Research, Hummingbird Technologies, Precision Decisions and the Science & Technology Facilities Council (STFC) Hartree Centre.

During the 24 month project, data such as soil type, crops and field operations will be collected by farmers and entered into Farmplan’s Gatekeeper software, which enables growers to supply field data without any duplication.

Gatekeeper then passes the data to aiScope which is based on IBM PAIRS, a geospatial-temporal information service. aiScope contains algorithms which combine the farmer’s data with additional data, such as weather, to help scientists create evidence-based blackgrass forecasting models to accurately predict where the weed will grow. Once farmers know where the blackgrass is likely to grow they can eradicate it in advance.

“Blackgrass continues to be a scourge of British farming and we are delighted to provide support to all those working hard to develop best practices to eradicate it,” said Alistair Knott, Proagrica’s Precision Agriculture Development Director. “At present, willing, enthusiastic growers across the UK have to supply data manually, when that data is already in their farm management system. They can now seamlessly and securely, expressly permission agreed historic data by field, in a standardised format, efficiently delivered via an API.”

“In order to develop forecast models and decision-support tools for a range of different crop threats, researchers need access to the highest quality data at scale. The aiScope project is working with Proagrica to make it easy for farmers to share their farm management data with researchers, allowing integration with environmental, survey, sensor and imaging data to develop more powerful models than ever before.  By applying the latest techniques in data science and artificial intelligence, we hope to address many different challenges, starting with Blackgrass,” said Blair Edwards, Technical Lead for the IBM Research Europe data science team in the UK.