Classification of Apple Types Using Principal Component Analysis and K-Nearest Neighbor

Abstract

Apple is a fruit that is quite popular in Indonesia and is widely consumed by people. This fruit has various types of shapes and colors. Types of apples can be distinguished by their color, size, and shape, but it is still difficult for ordinary people to type apples that are more similar in color and size, such as the examples of Braeburn and Crimson Snow apples. This gave rise to the idea of researching image processing to classify the types of apples. This is to help determine the differences between the two types of apples. The classification process of apples is done by testing the image of an apple based on existing training data. The research method consisted of preprocessing image segmentation with morphological operations and feature extraction into Principal Component Analysis (PCA). The classification algorithm used is a K-Nearest Neighbor (KNN). Using adequate training data will further improve the classification of types of apples. The final results of this study amounted to 91,67%.