DeepFruits: efficient citrus type classification using the CNN

Citrus fruit is a type of fruit with the same color or shape. This makes it difficult to distinguish between types of oranges. Various studies have proposed different techniques to deal with citrus classification. However, many approaches use conventional methods that often rely on operations based on visual abilities with drawbacks. This study aims to build a classification model using a convolutional neural network (CNN) to deal with problems in classifying citrus types quickly and accurately. To build a classification model, several stages such as collecting datasets, preprocessing, training, and testing datasets using several parameters to get the highest accuracy are required. Model testing is done using new data. The results show that this model can classify oranges with an accuracy rate of 96.0%. Some features of this study are the use of the dataset including diameter size, weight, and the average value of Red, Green, and Blue (RGB).