Automated Fruit Classification Based on Deep Learning Utilizing Yolov8

The act of fruit identification entails the discernment and classification of various fruit varieties, predicated upon their visual attributes. This task can be accomplished through various methods, including manual inspection, traditional computer vision techniques, and more advanced approaches using machine learning and deep learning. Our work recognized 15 types of fruit. Our experiment used the fruit imagery dataset, consisting of class Avocado, Banana, Cherry, Apple Braeburn, Apple golden 1, Apricot, Grape, Kiwi, Mango, Orange, Papaya, Peach, Pineapple, Pomegranate and Strawberry. Yolov8 is one of the most popular learning models in the field of object detection. Convolutional neural networks (CNNs), thus simultaneously predicting multiple bounding boxes and probability classes from these bounding boxes. Yolov8 trains the entire image and directly optimizes detection performance, modeled as a Convolutional neural network (CNN). The percentage of detection of test results using Yolov8n.pt for each of the 15 categories of fruit images with the average accuracy above is 0.97%, so it can be said to be quite good. After collecting data, Train training and system design implementation of Val Test Convolutional neural networks (CNN) algorithms enabled machine learning and were successfully detected using yolov8n. During the experimental process, we employ 100 epochs with the use of Google Collaboratory. Google Collaboratory has provided GPUs up to 12GB, making it easier and speeding up the fruit detection process.