Classification on passion fruit's ripeness using K-means clustering and artificial neural network

Abstract

This purpose of this study is to identify the level of ripeness of the passion fruit. The levels are classified into three distinguished stages: fruit in a ripe stage, a nearly ripe stage, and an unripe stage. The passion fruit-sorting system with artificial intelligence is an innovation of fruit sorting technology for industrial markets because it is very cost efficient and effective for a large production process instead of relying on manual labor process. The method used in this research is K-Means Clustering to perform passion fruit segmentation and Artificial Neural Network for classification based on RGB and A features. The input data is passion fruit video from 6 different sides. This study uses 75 passion fruit videos as training data and 20 videos as data testing with duration 5 seconds per video. The result achieves system accuracy of 90% with classification errors occur in the nearly ripe and unripe fruit due to the color closeness.