AI identifies crops from space with 90% accuracy

• Agrimetrics launches CropLens AI, a proprietary algorithm that identifies crop types from space.
• Identifications occur earlier in the season and with higher accuracy than previously possible.
• Commodity markets, supply chain forecasting, and farm input optimisation all stand to benefit.

Agrimetrics, the Agrifood Data Marketplace, has launched CropLens AI, an advanced machine learning algorithm that identifies crops from satellite-derived cloud-penetrating radar. This capability is a critical enabler for a variety of solutions, including disease and pest prediction, optimising nitrogen applications, irrigation, and yield prediction.

‍

Artificial Intelligence (AI) has been used to classify crops from space before. However, Agrimetrics CropLens AI is differentiated by:

1. An industry-leading max accuracy of over 90%, achieved in the one to two months prior to harvest.
2. High levels of accuracy earlier in the season, e.g. winter wheat can be predicted with 65% accuracy as early as October.
3. Improved usability and actionability through pre-linking to billions of relevant agricultural datapoints on the Agrifood Data Marketplace.
4. The use of cloud penetrating radar data, meaning the classifications are insensitive to cloud cover.

‍

Agrimetrics CropLens AI identifies 5 crop types – winter wheat, winter barley, oilseed rape (OSR), grass, and other – in UK fields; however, there are plans to expand this to include a wider range of crops and geographies. Historic data are available from 2017 and in-season predictions available now.

‍

“Real-time, accurate and reliable knowledge of what’s happening on the ground – any why – is now a critical success factor for actors across the agrifood value chain.” says Dr Matthew Smith, Chief Product Officer at Agrimetrics. “Agrimetrics CropLens AI and the Agrifood Data Marketplace are making it easier, faster, and more affordable to generate actionable insights earlier in the season.”

‍

Find out more about Agrimetrics CropLens AI, here.

‍

Agri-data ecosystem makes data more actionable

‍

Agrimetrics CropLens AI is one of a growing number of products available through Agrimetrics’s Data Marketplace. Airbus also recently confirmed Agrimetrics as an exclusive reseller of their entire earth observation portfolio for the UK agricultural market. This includes Airbus Crop Analytics, an advanced set of crop/field analytics that includes leaf area index, leaf water content, and soil water saturation.

‍

ClearSky, a cutting-edge AI outfit from the University of Hertfordshire, use the Data Marketplace to market uninterrupted (cloud-free) optical data from which measures such as NDVI can be calculated. Agrimetrics have also developed methods for identifying field boundaries from space and an accelerated SAR pipeline, that cuts access times from months to minutes. Other contributors include Centre for Ecology and Hydrology (CEH), Environmental Agency, and Weather Logistics.

‍

“Connecting data through our Data Marketplace provides huge benefits for users and data owners,” continues Matthew. “For users, we’re reducing costs, complexity, and time by providing more of the data they need in an analysis-ready format. For data owners, we’re providing them with a way to profit from their data without losing control of it.”

‍

A recent example of this was BASF’s wHen2gO, a decision support tool for oilseed rape herbicide applications. The application brought together several data sources – e.g. topography, cultivation, recent weather, soil type – to calculate the probability of herbicides leaching or running-off into waterways.

‍

Louis Wells, BASF Solutions Manager, said: “Agrimetrics brought all of the data together in one place. This meant that we didn’t have to engage with 5 or 6 different partners, meaning we could complete wHen2gO faster.”

‍