Development of Watermelon Ripeness Grading System Based on Colour Histogram

Abstract

In general, there are three common types of melon in Malaysia namely as watermelon, rockmelon, and honeydew but the most popular is the watermelon. Although melons are easy to find in the market, one common problem that often arises is how to choose a ripe melon with sweet and juicy taste. In this paper, an image processing procedure based on colour feature extraction approach for classification of watermelon ripeness was presented. The RGB value collected from the colour image histogram was used to determine ripeness grading of watermelon based on three qualities namely; ripe, overripe and underripe. A simple watermelon ripeness grading system was developed using GUI in MATLAB to ease the ripeness classification process. The statistical analysis shows that the system was 100 % accurately classified the ripe and overripe watermelon whereas for underripe, the accuracy was 96.3 %. Therefore, this proposed technique can reliably classified the ripeness of watermelon. Future work may include the implementation of data mining process for feature extraction using neural network or machine learning.