Page 859 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

In 2021, tomatoes accounted for more than 30% of Tanzania’s total agricultural output, yet these
diseases resulted in crop losses of up to 40% annually in the Southern Highlands. Smallholder
farmers, who make up a significant portion of the agricultural workforce, are especially
vulnerable. Traditional manual inspection methods are slow, imprecise, and inadequate for Agriculture 4.11: Smart
early intervention, underscoring the urgent need for a real-time, AI-powered disease detection
system. This solution aims to focus on the development of an Edge AI and IoT-powered
system for early tomato disease detection and smart control. The system integrates advanced
technologies to deliver accurate, real-time identification and notification of tomato leaf diseases.

Specific Objectives:
• To design and simulate a deep learning model capable of detecting and distinguishing
between healthy and diseased tomato leaves.
• To develop a real-time alert system that notifies farmers when disease symptoms are
detected.
• To build a mobile application that delivers insights and guidance based on model
outputs.
At the core of this solution is the YOLOv9 deep learning model, optimized for deployment on
edge devices such as Raspberry Pi. This architecture allows for real-time image analysis directly
on the farm, removing the dependency on continuous internet access an essential feature for
rural agricultural environments.

By enabling early and accurate disease detection, this AI-IoT solution supports smallholder
farmers in increasing crop yields, reducing losses, and minimizing unnecessary pesticide use.
The real time alerts and mobile based insights ensure that even farmers in remote areas can
take timely action, ultimately enhancing food security, promoting sustainable agriculture, and
improving farmer livelihoods through precision farming.

2�2 Benefits of the use case

Improve smallholder farmers’ livelihoods by reducing crop losses caused by tomato diseases.
Through early detection and targeted disease management, the solution minimizes income
loss and supports more stable, resilient agricultural practices.

Strengthen food security by increasing tomato crop yields and ensuring a more reliable
food supply. By enabling timely and accurate disease intervention, the system contributes to
sustained agricultural productivity and helps combat hunger in rural communities.

Support ecosystem conservation and biodiversity protection by promoting sustainable disease
control techniques. The reduction in unnecessary pesticide use helps maintain soil health and
reduces harm to beneficial organisms, aligning with broader goals of sustainable land use and
environmental stewardship.

2�3 Future Work

Future work includes expanding the dataset to cover a wider range of tomato varieties and
environmental conditions to improve model robustness. Integrating drone based monitoring
is planned to enhance large scale field surveillance and complement sensor based ground
monitoring systems.

823