Open AccessIdentifikasi Citra Digital Jenis Beras Menggunakan Metode Anfis dan Sobel
Author(s) -
Gansar Suwanto,
Riza Ibnu Adam,
Garno
Publication year - 2021
Publication title -
jurnal informatika polinema
Language(s) - English
Resource type - Journals
eISSN - 2614-6371
pISSN - 2407-070X
DOI - 10.33795/jip.v7i2.406
Subject(s) - sobel operator , artificial intelligence , grayscale , adaptive neuro fuzzy inference system , computer vision , pattern recognition (psychology) , computer science , digital image , texture (cosmology) , fuzzy inference system , image (mathematics) , image processing , mathematics , edge detection , fuzzy logic , fuzzy control system
Rice is one of the leading national food products and superior agricultural products in Indonesia. The many types of rice in Indonesia make it increasingly difficult to distinguish rice by just relying on the eye. Because each type of rice has relatively different shape and texture characteristics. Therefore, digital images can be used as a first step in identifying types of rice. This study aims to identify the types of rice using image processing. Taking the value of the shape characteristics using the morphology method and compared with the sobel method. While taking the value of the texture features using the grayscale image method. Then, the value of the shape and texture do the grouping according to the type of rice. The data used in this study were 140 images. 100 of the 140 images were conducted training using the ANFIS (Adaptive Neuro Fuzzy Inference System) method by utilizing the value of the shape and texture of the image. The test was carried out 5 times using 140 images. The test results using the ANFIS (Adaptive Neuro Fuzzy Inference System) method by 85.2%. Meanwhile, sobel edge detection can affect accuracy by 3%.