Fruit Classification Using Traditional Machine Learning and Deep Learning Approach

Advancement in image processing techniques and automation in industrial sector urge its usage in almost all the fields. Fruit classification and grading with its image still remain a challenging task. Fruit classification can be used to perform the sorting and grading process automatically. A traditional method for fruits classification is manual sorting which is time consuming and involves human presence always. Automated sorting process can be used to implement Smart Fresh Park. In this paper, various methods used for fruit classification have experimented. Different fruits considered for classification are five categories of apple, banana, orange and pomegranate. Results were compared by applying the fruit-360 dataset between typical machine learning and deep learning algorithms. To apply machine learning algorithms, basic features of the fruit like the color (RGB Color space), size, height and width were extracted from its image. Traditional machine learning algorithms KNN and SVM were applied over the extracted features. The result shows that using Convolutional Neural Network (CNN) gives a promising result than traditional machine learning algorithms.