Misdiagnosing maize diseases causes significant losses for Ghanaian farmers. Learn how Aya enabled rapid disease identification through a smartphone app.
Maize is a crucial crop for Ghana’s agriculture, but diseases affecting maize plants remain a significant problem for farmers. Identifying and treating these diseases correctly is challenging because many maize diseases look similar, especially for farmers without advanced training. Agronomists who can diagnose these diseases are rare and costly in rural areas, making it hard for farmers to get the help they need.
Aya Data partnered with Demeter Ghana to tackle this issue with a pro-bono project. We aimed to develop an easy-to-use smartphone app that could help farmers quickly and accurately identify maize plant diseases. Here’s how we did it:
1. Data Collection: We collected 5,000 images of maize plants, both healthy and diseased, using smartphones. Agronomy experts helped label these images, noting the type and severity of each disease.
2. Model Development: Using these labelled images, we trained an Artificial Neural Network (ANN), a type of AI that can learn to recognize patterns in data. We then integrated this model into a simple smartphone app.
3. App Deployment: The app was distributed to farmers, allowing them to take and submit photos of their maize plants for disease diagnosis.
The app achieved impressive results:
- High Accuracy: The AI model could accurately identify disease types and their severity with 96% accuracy.
- Empowering Farmers: The app provided farmers with a valuable tool to quickly diagnose and address plant diseases, helping them manage their crops more effectively and improve yields.
This initiative helped Demeter Ghana support farmers by using cutting-edge technology to solve real-world problems in agriculture.
Interested in enhancing your crop management with AI?
Contact us today to learn how our disease detection models can help optimise your farming practices and increase yields.
Disclaimer: Aya Data respects client confidentiality and will not disclose any specific details about clients or projects. Any identifying information in our case studies may be anonymized.