Title: Project Title: Raspberry-Pi Based System For Real Time Poultry Egg Quality Assessment Using Convolution Neural Network. An Affordable Computer Vision Solution for Small-Scale Poultry Farmers in Tanzania
Authors: Isaka Sabale Makoye, Agnes Simon Hamisi, Alex Anjelo Mkunangwa, Dr. Danny Mfungo
Volume: 10
Issue: 4
Pages: 237-242
Publication Date: 2026/04/28
Abstract:
The poultry industry in Tanzania plays a crucial role in food security and economic stability, yet it remains hindered by inefficient manual methods for egg quality assessment. Traditional techniques such as candling are labor-intensive, prone to human error, and economically unsustainable for small to medium-scale farmers. To address these challenges, this project proposes the development of EggFresh AI an intelligent, computer vision-based system for the automated detection and sorting of rotten eggs in real-time. The system utilizes a Raspberry Pi microcomputer integrated with a high-resolution camera, LED lighting array, and servo-controlled sorting mechanism. By applying machine learning algorithms, specifically a Convolutional Neural Network (CNN) model trained on locally sourced egg images, the system classifies eggs as fresh, questionable, or rotten with a target accuracy exceeding 95%[1]. A user-friendly web dashboard provides real-time monitoring and analytics, enabling farmers to track quality metrics and system performance. This project emphasizes affordability, scalability, and local relevance, aiming to reduce post harvest losses, enhance food safety, and improve operational efficiency for Tanzanian poultry producers. By leveraging accessible technology and adaptive AI, EggFresh AI seeks to bridge the gap between high-cost industrial automation and the practical needs of local farmers, contributing to the digital transformation of agriculture in Tanzania.