Development of a Fusion-Based Method to Enhance YOLO Precision for Edward Mango Ripeness Detection via Gaussian-Weighted Vibrational Modeling in IQF Automation

This paper presents the implementation of a computer vision system fused with a vibrational model to classify Edward mangoes by ripeness in an automated IQF environment. The system replaces subjective manual selection with reliable analysis of external features-surface color, shape, and firmness-routing mangoes to reject or process. Classified mangoes undergo automated cleaning via immersion in alkaline and neutral detergents at 60°C before IQF. Controlled by a PLC, the system synchronizes transport, selection, and washing in a compact, cost-effective design tailored for Edward mangoes, processing up to 1080 fruits per hour, achieving 95% accuracy in ripeness classification. This implementation reduces labor dependency by 40%, enhances product consistency, and improves post-harvest efficiency by combining two validation parameters in one system.