UK-Brazil-Ghana Partnership for cassava disease prevention

Early disease prediction for cassava crops

This collaboration between ScareCrow Intelligence (UK), IA Sense (Brazil), and the Council for Scientific and Industrial Research (CSIR) (Ghana) will develop an AI-powered early prediction of cassava fungal diseases.

The project addresses critical food security challenges by combining environmental monitoring with computer vision techniques to provide early disease prediction for cassava crops. Cassava, a vital staple feeding over 800 million people globally (FAO, 2013), serves as the primary calorie source for millions across Africa, Latin America, and Asia. However, fungal diseases including anthracnose, root rot, and blight pose severe threats to production. These diseases can devastate entire harvests, causing yield losses of 20-100% (Bandyopadhyay et al., 2006) and contributing to food scarcity that affects millions of vulnerable households.

How the prediction system works

Current reactive management practices lead to excessive fungicide use, with farmers applying chemicals every 10-14 days regardless of actual disease pressure, resulting in environmental damage and unnecessary costs. The proposed system integrates environmental monitoring with computer vision techniques specifically developed for local conditions. By analysing environmental data such as air temperature and humidity, and visual symptoms, the new system predicts fungal outbreaks days in advance. This early warning capability allows targeted interventions only when necessary, drastically reducing fungicide applications while improving disease control effectiveness.

The project scope encompasses four phases. First, they will identify and prioritise target fungal diseases based on their environmental triggers, economic impact, and prevalence in Ghanaian cassava farming. Second, they will establish a data collection framework combining historical records, real-time environmental monitoring, and field images. Third, they will develop and enhance predictive models using Large Language Model (LLM)-powered AI to achieve high accuracy. Unlike legacy models that rely on historical patterns, their AI continuously learns and adapts, ensuring earlier and more precise warnings. Finally, they will validate system performance through controlled laboratory testing before conducting field pilots, ensuring the technology delivers reliable results under real-world farming conditions.

Meeting development goals

The project contributes to multiple UN Sustainable Development Goals. For SDG 2 (Zero Hunger), they will help improve food security by reducing crop losses that currently threaten nutrition for millions. For SDG 13 (Climate Action), they will help reduce chemical inputs while building climate resilience through adaptive disease management. The system also supports SDG 12 (Responsible Consumption) by optimising resource use and SDG 15 (Life on Land) by protecting biodiversity from chemical contamination.

For more information

For more information on this project, contact us, or view all projects funded under the Climate-Smart Agriculture Partnership programme.

Innovate UK Climate-Smart Agriculture Partnership: UK-Brazil-Africa brings together innovative people and organisations to promote climate-smart agriculture in Africa.