Analisis Tingkat Kematangan Buah Jeruk Menggunakan Chain Code Dan KNN (K-Nearest Neighbors) Berbasis Website

Abstract

This research aims to develop a web-based orange ripeness classification system using the K-Nearest Neighbor (KNN) method, leveraging morphological and color features as the main parameters. The image processing workflow begins with converting RGB images into the HSV color space, followed by object segmentation using the thresholding method, and feature extraction including chain code, area, and shape factor. The dataset consists of 50 orange images as training data and 20 orange images as test data. The evaluation was conducted in two scenarios: single testing and batch testing. The single testing on 5 test images achieved a perfect classification accuracy of 100%. In batch testing, the system achieved an accuracy of 0.90. These results indicate that the system is capable of effectively classifying orange ripeness, with a very low rate of false-positive predictions. The application is implemented as a web-based platform, making it easily accessible, and is expected to serve as a practical tool for sorting and grading oranges based on their ripeness levels.

Keywords: Orange Classification, K-Nearest Neighbor, Feature Extraction, Image Processing, Web Application.