Identifikasi Varietas Durio Zibethinus Berdasarkan Sebaran Trikoma Daun Menggunakan GLCM dan KNN

Abstract

Durio zibethinus known as Durian is a family of Bombacaceae, that can also be categorized into Malvaceae with genus name Durio. The objectives of this research is to apply the Gray Level Co-occurrence Matrix (GLCM) feature extraction in trichomes image of durian leaf and identify 10 varieties of Durio zibethinus. This image si captured using microscope camera digital Dino Lite. Trichomes image is captured with 200 times magnification and produce 640x480 pixels of image with BMP format. The initial process is to change Red, Green, Blue (RGB) image into grayscale, followed by performing feature extraction with GLCM. K-fold cross validation with 5-fold is used to divide the training data and testing data. Classification is performed using the K-Nearest Neighbor (KNN). It is found that the highest accuracy is 78% at 3 clusters, GLCM's angle 90° and 1 pixel distance