DESIGN AND FABRICATION OF AN AI ALGORITHM OF A DRONE-BASED METHANE EMISSION ANALYSIS FOR ENVIRONMENTAL MONITORING AND FUTURE RELEASE PREDICTION

Methane (CH₄) is a potent greenhouse gas with a global warming potential over 25 times higher than that of carbon dioxide. In Nigeria, unmonitored methane emissions from livestock, waste, and oil activities significantly contribute to climate change and local air pollution. This study presents the design, fabrication, and testing of a low-cost, AI-assisted methane sensing system integrating an MQ-4 sensor module, Arduino microcontroller, and drone-based deployment for real-time methane detection and emission prediction. The fabricated system was field-tested at the Mubi Cattle Market, Adamawa State a representative methane emission site due to livestock waste accumulation and organic decomposition. Data from the sensor were transmitted via Global System for Mobile Communications (GSM) to a cloud-based platform, processed using machine learning algorithms (Random Forest and Long Short-Term Memory LSTM models) to identify emission trends and predict short-term variations. The system achieved a detection sensitivity of 92.5% and a correlation coefficient (R²) of 0.94 against reference analyzer readings. Results demonstrate the potential of combining low-cost sensing, drone mobility, and AI for scalable methane monitoring in developing regions. The project contributes to local technological capacity and supports Nigeria’s national climate commitments under the Sustainable Development Goal (Climate Action).