Classification of Banana Types Using the Convolutional Neural Network

Banana is a fruit that is quite familiar because it is easy to reach, in general, this fruit is consumed for dessert, and can be made for processed foods such as banana cakes, banana chips, and other preparations using a mixture of bananas and is also beneficial for body health. Many types of bananas are found in Indonesia, such as Barangan bananas, Banten bananas, Raja bananas, Ambon bananas, horn bananas, Mas bananas, Kepok bananas, and other types. The many types of bananas make it difficult to distinguish them, generally to recognize bananas is from skin color and size alone. Therefore, digital image processing technology is needed to be able to recognize and classify bananas in a fast and precise time. In this study, 4 types of bananas will be classified, namely Ambon bananas, Barangan bananas, Raja bananas, and Mas bananas. The method used in this research is Convolutional Neural Network. The data used are 1200 images and divided into training data of 960 images and test data of 240 images. After testing, it can be concluded that the method used is quite good in classifying 4 types of bananas and gets an accuracy of 93.75%.