Identification of Deterioration Areas in Green Vegetables Using Image Segmentation Algorithm Based on Green Vegetable Photos

This research is in the field of image processing, particularly image segmentation, aimed at identifying deterioration areas in green vegetables based on photographic images. The study introduces a novel combination of K-Means clustering for color segmentation and GLCM (Gray-Level Cooccurrence Matrix) for texture analysis to detect freshness levels in vegetable images. A total of 200 training images and 100 testing images—classified into fresh and non-fresh categories—were used in this study. The preprocessing stage included noise reduction, contrast enhancement, and normalization. The results demonstrated high accuracy, achieving 96% in correctly classifying vegetable freshness. The novelty of this study lies in the integration of unsupervised color clustering and second-order texture statistics for local, image-based detection of vegetable quality. This approach presents a scalable, low-cost alternative to manual inspection, with potential applications in smart retail environments and future integration with IoT-based real-time monitoring systems to support sustainable food supply chains. This method contributes to the development of intelligent decision-support systems for improving efficiency and reliability in agricultural product management. Further research can focus on extending the model to various vegetable types and deploying it on edge computing platforms for real-time retail usage.