LAIRG - AI for Cattle Counting & GHG Emissions

Overview

A Makerere University-led project is deploying UAV imagery and real-time AI to count cattle, classify farming systems, and estimate greenhouse gas emissions across Mubende District giving Uganda its first credible, data-driven foundation for climate action in the livestock sector.

Uganda's livestock sector is a major but poorly understood contributor to the country's greenhouse gas emissions. Traditional manual cattle surveys are slow, expensive, and incomplete leaving policymakers without the data they need to design effective climate interventions. In a country where livestock farming underpins millions of livelihoods, the absence of reliable emission estimates has made it nearly impossible to meet national and continental climate commitments.

The project uses drone-captured video and images processed by advanced AI algorithms YOLO V4 for detection and SORT for tracking to detect, count, and classify cattle in near real-time. The system automatically estimates GHG emissions by farming system type, generating the granular, scalable data that manual methods cannot. Built on open-source code and developed in partnership with local communities through a Participatory Extension Approach, the solution is designed to be transparent, reproducible, and owned by the people it serves.

The project directly advances Uganda's National Development Plan III, the African Union's Agenda 2063, and SDG 13 on Climate Action providing the evidence base for targeted livestock emission reduction policies. Beyond data, it is building a new generation of AI-capable researchers: with gender-inclusive training requirements embedded from the start, the project ensures that the benefits of AI in agriculture are shared equitably. As the model scales, it offers a replicable framework for livestock monitoring across sub-Saharan Africa, where climate-smart agriculture is urgently needed but data remains scarce.