Unified Color Space Segmentation and Yolov5 Approach for Nutmeg Ripeness Detection and Quantity Estimation

The vital aspect in maintaining consistent quality across the global spice sector is the accurate analysis of nutmeg maturity and harvest volume. The evaluation accuracy degrades with the use of the traditional methods which causes inconsistency and reduced effectiveness. To address the issues posed by the conventional methods, this work proposes a novel AIpowered system that integrates color space segmentation and YOLOv5 detection model. In order to upgrade feature extraction, intrinsic color variations are leveraged by the segmentation method. To accomplish fast, on-the-fly detection and counting, YOLOv5 is utilized. The empirical results demonstrate the efficacy of the system compared to the traditional methods by achieving 95% accuracy, 94% precision and 90% recall. This study serves as a reliable and scalable approach to nutmeg assessment with the scope for improvements to handle overlapping fruits by leveraging advanced image processing methods.