Strawberry ripeness classification system based on skin tone color using multi-class support vector machine

Abstract

This research aims to build an automatic sorting system for strawberry ripeness into three categories: unripe, partially ripe, and ripe. Manual fruit sorting has many weaknesses and limitations. One of the disadvantages is human error in the sorting process. Therefore, the implementation of artificial intelligence as replacement of human worker can mitigate the problem. Fruit ripeness is identified based on color characteristic, which is the Red, Green, Blue (RGB) value of the object. Multi-Class Support Vector Machine (SVM) with Radial Basis Function (RBF) kernel function is implemented to classify the ripeness. The data was taken using the Logitech C920 web camera which then divided into training and testing data video. In this research, prototype of strawberries sorting is built with real-time video which is never been considered in previous researches. Training data consists of 70 unripe strawberries, 70 partially ripe strawberries, and 70 ripe strawberries. Meanwhile testing data comprises of 30 unripe strawberries, 30 partially ripe strawberries, and 30 ripe strawberries. The result shows that the strawberry ripeness classification system using Multi-class SVM with RBF kernel function produces up to 85.64% accuracy, where the parameters are C = 7 and gamma (γ) = 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-2</sup> .

Access to Document

Other files and links

• Link to publication in Scopus

Fingerprint