Segmentasi Tomat Menggunakan Metode K-Means Clustering dan Pengolahan Citra Digital

Abstract

The implementation of image processing in the plantation sector has been widely researched and developed, for example, to identify fruit maturity and control fruit harvesting robots. These systems need the primary process, namely segmentation, to determine the fruit area and background. This study aims to apply the tomato segmentation method. This method consists of four main processes: region of interest (ROI) detection, pre-processing, segmentation, and segmentation. Resizing and thresholding with the Otsu method were applied to ROI detection. RGB to HSV color space conversion was used in the pre-processing. Next, K-means clustering is applied to the segmentation, followed by implementing morphological operations to remove the remaining noise. Evaluation of the performance of the tomato segmentation method on 160 images showed that the average precision, recall, and F-measure values obtained were 97%, 88%, and 94%, respectively.