AI for Tunisian Farmers – Predicting Date Diseases Caused by Oligonychus afrasiaticus

1. Introduction: Why This Project?

Date palm cultivation is a cornerstone of Tunisian agriculture, particularly in southern regions like Kébili and Tozeur. Tunisia is one of the world’s leading exporters of the Deglet Nour variety, renowned for its high quality. However, the industry faces several threats, including pests and diseases, one of the most damaging being Oligonychus afrasiaticus.

This mite, commonly known as the dust mite of dates, attacks the fruit, causing it to dry out and significantly reducing its market value. Currently, detection relies mainly on visual inspection, which is often too late to prevent substantial losses.

Our project aims to develop an AI-powered solution to help Tunisian farmers detect and predict infestations of Oligonychus afrasiaticus early, ensuring better protection of their crops.

A Strategic Partnership for In-Depth Expertise

To ensure the reliability of our solution, we are fortunate to collaborate with a partner specializing in agriculture and crop protection. This expert partner helps us:

Identify the early signs of Oligonychus afrasiaticus infestations.
Define the best practices for controlling this pest.
Access reliable field data to train our AI model.

Thanks to this collaboration, our project combines scientific expertise and cutting-edge technology to offer an effective and field-adapted solution.

2. Problem Definition

Why Is This a Problem?

Rapid Spread: Oligonychus afrasiaticus thrives in hot and dry climates, which are common in Tunisia.
Crop Damage: The mite sucks the sap from dates, causing premature drying and discoloration, making the fruit unsellable.
Inefficient Detection Methods: Farmers must manually inspect date palms, which is time-consuming and often inaccurate.
Economic Losses: Undetected infestations lead to a significant drop in yield and quality, directly affecting farmers' incomes.

3. Challenges and Justification for AI Use

Why Can AI Help?

Artificial intelligence can analyze images of dates to detect early signs of Oligonychus afrasiaticus infestation. Using a machine learning model, we aim to:

Identify infected dates based on their external appearance.
Predict infestation trends based on climate conditions and historical data.
Alert farmers in real-time and provide preventive recommendations.

Impact on Tunisian Farmers

Reduced losses through early detection and quick intervention.
Optimized pesticide use, preventing overuse and environmental harm.
Improved date quality, increasing their market value and farmer revenues.

4. Approach and Methodology (CBL - Challenge-Based Learning)

We follow the Challenge-Based Learning (CBL) methodology, which consists of three phases: Engage, Investigate, and Act.
We are currently in the first phase:

Phase 1 – Engage: Identifying and Understanding the Problem

🔹 Big Idea:
Agriculture

🔹 Essential Question:
How can we reduce the impact of Oligonychus afrasiaticus (السَديَّة) on dates and effectively control their spread?

🔹 Challenge:
Protect date palms from Oligonychus afrasiaticus.

This phase allows us to gain a clear understanding of the issue and lay the foundation for exploring AI-based solutions.

📝 PV of the 2nd Week - Artificial Intelligence Project (CBL Methodology)

Date: 31/01/2025
Location: Google Meet

👥 Attendees:

Aziza Neffati (Moderator)
Amine Mnif (Technologist)
Yassine Hemissi (Time Keeper)
Ibtissem Ayedi (Technologist)
Islem Ben Mansour (Note Keeper)
Amal Ben Amor (Speaker)
Khouloud Kechim (Devil’s Advocate)

1️⃣ Opening of the Meeting

Time: 07:00 PM
We recalled our main idea, which is to develop an AI-based solution for managing and preventing insect infestations in forests.

2️⃣ Defining the Engagement Phase

We determined the engagement phase with the following framework:

Big Idea:

Agriculture

Essential Question:

How can we reduce the impact of mites on dates and effectively combat their proliferation?

Challenge:

Minimize the impact and spread of mites while preserving the quality of dates.

3️⃣ Presentation of AI Project Ideas

We discussed the identification of our partner and the essential information they provided:

Presence of other insects, such as dragonflies.
Mites do not have a specific product for their eradication.
If 50% of the forest is affected by this insect, it is necessary to use a product on the entire forest.

4️⃣ Call-to-Action Video

We discussed the creation of a Call-to-Action video:

Developing the script.
Choosing the appropriate editing tools.

5️⃣ Presentation Template

We selected the template for our presentation for validation.

6️⃣ Documentary Blog

The structure for the documentary blog was also chosen and validated.

Thank you for being part of our journey!