Machine Learning as a Strategic Tool for Helping Cocoa Farmers in Côte D'Ivoire

Area of Science:

• Environmental Science
• Computer Science
• Agricultural Science

Background:

• Climate change poses significant threats to agriculture, impacting crop yields and water availability.
• Developing countries face unique challenges in adopting advanced agricultural technologies due to cost constraints.
• Smart agriculture, powered by artificial intelligence (AI), offers potential solutions for environmental sustainability and economic development.

Purpose of the Study:

• To explore the application of AI and machine learning for social good in agriculture, focusing on cost-effective solutions for developing countries.
• To develop and evaluate machine learning models for two key agricultural tasks: crop health monitoring and water resource management.
• To demonstrate the feasibility of using low-cost and open-source tools for AI implementation in smart agriculture.

Main Methods:

• Deep neural networks (YOLOv5m) were employed to detect healthy and damaged cocoa plants and pods using mobile phone imagery.
• Remote sensing data from NASA's GRACE Mission and ERA5 (Copernicus Climate Change Service) were analyzed for water management.
• A novel deep neural network architecture (CIWA-net) with a U-Net-like structure was proposed for forecasting total water storage anomalies.

Main Results:

• The YOLOv5m model successfully distinguished between healthy and damaged cocoa plants and pods from mobile phone images.
• The CIWA-net architecture demonstrated potential in forecasting total water storage anomalies.
• The study confirmed the viability of using accessible hardware, software, and open data for AI in agriculture.

Conclusions:

• AI and machine learning, particularly with low-cost tools, can empower small farmers and regional stakeholders to implement effective climate change countermeasures.
• The developed models show promise for improving agricultural cultivation practices and enhancing water resource management in developing regions.
• This research highlights the potential of AI to contribute to environmental sustainability and economic resilience in the agricultural sector.