Identifikasi Tingkat Roasting Biji Kopi Menggunakan Algoritma K-Nearest Neighbors (KNN) pada Platform Mobile

Abstract

Coffee is a tropical plant that also serves as a non-alcoholic beverage containing caffeine. There are many benefits of consuming coffee, including its caffeine content that can increase the body's metabolic rate. For some individuals with nighttime routines, coffee can be a good alternative beverage as the caffeine in it helps overcome drowsiness. Additionally, coffee has good antibacterial properties that can aid in the treatment of various health issues. This research aims to develop a mobile application that can identify the roasting levels of coffee beans using the K-Nearest Neighbors (KNN) algorithm. The HSV (Hue-Saturation-Value) image processing method is used as a pre-processing step in this application. The application is developed using the Flutter framework with support from the OpenCV library. Furthermore, the application utilizes Flask as an API to connect the mobile application with the backend server. Testing is conducted using a dataset of coffee beans with various roasting levels that have been collected and processed beforehand. The test results demonstrate that the KNN algorithm is capable of identifying the roasting levels of coffee beans with significant accuracy. This application provides convenience for coffee enthusiasts to quickly and practically recognize the roasting levels of coffee beans through their mobile devices. This research is expected to contribute to the development of coffee bean recognition technology and its broad application in the coffee industry.