<![CDATA[Manual egg grading and fragmented supply chain management in poultry SMEs often cause inconsistent egg quality decisions, inaccurate inventory records, and avoidable distribution losses. This study aims to develop and evaluate an integrated supply chain management system that embeds computer vision–based egg quality detection to improve real-time operational control and decision-making. The system is implemented as a digital SCM application connected to a convolutional neural network model trained on 1,200 labeled egg images across five quality categories (good, cracked, dirty, fertile, non-fertile), with performance assessed using accuracy, precision, recall, F1-score, and inference time, alongside before–after operational and economic measurements in an SME workflow. Results show the CNN achieves 92% validation accuracy with average precision 0.92, recall 0.94, F1-score 0.93, and 0.11 seconds per egg inference, enabling practical real-time classification. After integration, egg sorting accuracy increases from 75.5% to 90.2%, inspection time decreases by 81.3%, sorting capacity rises 5.3×, and inventory accuracy improves from 82% to 98%, reducing daily stock discrepancies by 85%. The novelty lies in tightly coupling computer vision quality outputs with SCM inventory and distribution modules, creating immediate stock updates and automated control points. The findings imply that AI-enabled digital supply chain management strengthens quality assurance, inventory optimization, and SME profitability, supporting scalable modernization of food supply chains through deployable computer vision systems. Highlights: Integrated computer vision raises egg sorting accuracy from 75.5% to 90.2%. Real-time CNN classification reaches 92% validation accuracy with 0.11 s inference. Digital SCM lifts inventory accuracy to 98% and delivers 193% ROI payback. Keywords: Computer Visio, Supply Chain Management, Egg Quality Detection, Convolutional Neural Network, Poultry SMEs, Inventory Optimization]]>
Copyrights © 2026
