Segmentasi Tingkat Kematangan Buah Pisang Cavendish Sangat Matang Berdasarkan Warna Menggunakan Watershed

Abstract

The Cavendish banana is considered to be one of the most sought-after export-oriented fruit in Indonesia. The large market size of export fruits, especially Cavendish bananas, opens opportunities for Indonesia to raise both its product quantity and quality in order to increase its competitiveness. Various methods have been conducted to increase the quantity and quality of Indonesia's cavendish bananas, one of which is the adoption of the image processing technology. This method aims to simplify and resolve cultivation and processing issues regarding Cavendish bananas, among other things by minimizing human error in determining fruit ripeness (which is traditionally conducted by manual labor). This research uses HSV segmentation and a Watershed algorithm in segmenting images of 50 ripe Cavendish bananas with 40 training data and 10 testing data. Based on our research, we founda out that ripe Cavendish bananas have 42% red, 37% green, and 21% blue in average, with an accuracy rate of 65%