PENGELOMPOKKAN TINGKAT KEMATANGAN BUAH JERUK MENGGUNAKAN METODE K-MEANS CLUSTERING

Abstract

Kabupaten Kampar is one of the citrus cultivation centers in Riau, the most famous orange is the Siamese orange from Kampar. Siamese oranges from Kampar have characteristics that are different from other oranges such as having a sweet taste, fragrant aroma, and thin skin. Currently, the fruit ripeness classification process still uses the manual method farmers make direct visual observations by looking at the shape, size and color of citrus fruits. This study aims to design a classification system for the maturity level of citrus fruits using the k-means clustering method. The data used are images of citrus fruits recorded using a smartphone camera and saved in JPG format that then extracts each RGB and uses as input data. The sample data used in this study amounted to 85 represented into three levels of ripeness, namely unripe, ripe and overripe fruit. A total of 70 data were used as training data and 15 data were used as testing data. In the training data, the system managed to recognize 19 fruits as unripe clusters, 42 fruits as ripe clusters and 9 fruits as overripe clusters. The level of accuracy in the training data is 91.42%.