Fruit Ripeness Classification System Using Convolutional Neural Network (CNN) Method

Abstract

The increasing consumer demand in the fruit industry has also demanded that various sectors of the fruit processing industry be able to adapt to this situation. The demand for good quality and fresh fruit requires technological advances and supporting systems that can be used in the fruit processing industry to produce the best quality fruit. Referring to this, this study aims to detect the type and maturity of fruit using machine learning with the CNN (Convolutional Neural Network) method using the function of a camera that is integrated with the program algorithm. This research is a refinement of previous research that has been made at the university by increasing the ability to read objects based on color with different methods. In this programming language, Python also requires several additional libraries to carry out the object detection process, namely by using the cvzone library as the main library. This study shows that the detection of fruit and ripeness using the CNN method was successful in detecting the type and maturity of the fruit. In the design and trial of this research, it can run well according to the algorithm created by the researcher. The success rate and accuracy of the detection of the type and maturity of this fruit reach 90%.