The challenge

Frequently published Department for Environment Food and Rural Affairs (DEFRA) reports continue to conclude that farm field run-off - from excessive chemical, fertiliser, manure, and bio-solid application is contributing to the deteriorating health of our nation’s river network.

While the aforementioned are frequently and intensely reported on programmes such as BBC Radio 4’s Farming Today, what often goes seldom reported are large-scale crop failures which decimate entire yields for farmers through conditions such as blight, mildew and fungal infections - where it is too late to apply preventative measures/treatments once the initial indicators are observed.

Simultaneously, the current conflict in Ukraine alongside the legacy of the coronavirus pandemic has resulted in “the greatest cost of living crisis” in over 50-years - with everyday staples such as fuel, grains, cooking oils and chemicals suffering from supply-chain issues and dramatic cost inflation. A further product, which has seen a >300% cost increase in 12-months, is agricultural fertiliser - vital in the UK’s (and global) sustainable food production.

As such a ‘precious commodity’, fertiliser (and other treatment) application must now be highly targeted amongst crops to avoid expensive wastage. Currently, the majority of agricultural monitoring and sensing technology is large, expensive, insensitive, and non-automated. However, the advent and growth of low-cost, highly miniaturised, Internet of Things (IoT) enabled, and Artificial Intelligence-fused sensor networks across other applications represents a unique ‘capability translation’ for enhancing modern agriculture and solving these endemic problems. 

What we're doing

Following our prior successful deployment of a radiation sensor network in the Chernobyl Exclusion Zone (ChEZ), this project will seek to refine and deploy further low-cost, resilient, and low power consumption modules of crop and environmental sensors into an agricultural environment.

Having developed prototype field-ready systems by extending our proven hardware and (cloud-based) software ‘backbone’, real-time changes/variations in pH, eH, dO2, crop/soil colour, fluorescence, PM2.5, water level, humidity, soil hardness, and nutrient/chemical composition will be captured by the array of in-field devices and automatically relayed to the Amazon Web Services (AWS) cloud. Once uploaded, our self-learning artificial intelligence (AI) and machine learning (ML) algorithm will be able to predict changes from this converging data before they occur - guiding farmers to best manage, intervene, optimise, and protect.

How it helps

This project will allow for farmers to optimise their utilisation of increasingly expensive supplements and treatments to (i) enhance yield and (ii) avert crop failures through the early identification of potential pests and diseases before it is otherwise too late. It will also prevent ‘surplus’ macro- and micro-nutrients being needlessly applied which typically run-off arable land and damage watercourses.

Investigators

• Dr Peter Martin, Physics
• Dr Freddie Russell-Pavier, Physics
• Andrew Hughes, Campus Division (UoB)
• Lucy Antysz, Amazon Web Services
• Dr Adam Land, DEFRA