<![CDATA[Green onions, commonly used in Indonesian cuisine, have significant agricultural potential. Despite high production, their quality, particularly freshness, is traditionally evaluated visually, leading to inconsistent and subjective results. This study aims to develop an objective and accurate method for classifying the freshness of green onions using an Artificial Neural Network (ANN). Previous studies have employed ANN but have not specifically targeted the freshness classification of leeks. The proposed method utilizes the color and texture features of green onions.The research methodology includes image acquisition, preprocessing, segmentation, morphology, feature extraction, and classification using ANN. A total of 300 images were acquired and categorized into three freshness levels: not fresh, less fresh, and fresh. During the training phase, 240 images were used, and 80 images were reserved for testing. The optimal feature combination identified includes HSV and LAB color features along with texture features (Contrast + Energy). The results demonstrated that the freshness classification of green onions achieved 100% accuracy in both training and testing phases. The training process, with 240 images, had a computation time of 142.684 seconds, while the testing process, with 80 images, took 35.648 seconds. These findings indicate that using ANN based on color and texture features is highly effective in determining the freshness level of green onions.]]>
Copyrights © 2024
