THE CLASSIFICATION OF IMAGE ORANGE OF KINTAMANI BASED ON THE COLORS AND SIZES BY EMPLOYING EUCLIDEAN DISTANCE APPROACH

Abstract

This research is aimed at: 1) Designing the classification system of orange
of Kintamani based on the colors and sizes by employing Euclidean Distance approach, 2)implementing the classification system of orange of Kintamani based on the colors and sizes by employing Euclidean Distance approach. They consist of 5 types of process namely color feature extraction, preprocessing operation (Grayscale, Thresholding), size feature extraction, Euclidean distance, and image classification. Testing image and training data image of this application are the input image with bitmap (*.bmp) extension and the outputs contain information about orange quality. Testing was done to all images of orange that become training data added with non training data of orange image. Process of testing requires farmers’ help to classify the orange before the image is taken.
In designing and implementing the application, waterfall method or classic life cycle model was employed. It belongs to classical and systematical model in developing software. It covers some stages: 1) requirements definition, 2) system and software design, 3) implementation and unit testing, 4) integration and system testing, and 5) operation and maintenance.
The implementation and the testing is a classification system of orange of
Kintamani that use Borland Delphi 7. programming language. From the
performance testing data, it was found that the system is able to identify the orange quality up to 98,33 % with 60 total testing samples of orange image of Kintamani. It shows that the system can be used to help the farmers in classifying orange quality of Kintamani.