Open AccessModel for identification of rice type using combination of shape and color features
Author(s) -
J. Jumi,
Achmad Zaenuddin,
Tedjo Mulyono
Publication year - 2021
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/1108/1/012038
Subject(s) - artificial intelligence , pattern recognition (psychology) , color histogram , mathematics , histogram equalization , grayscale , cluster analysis , color difference , feature extraction , weighting , feature (linguistics) , similarity (geometry) , histogram , computer science , computer vision , color image , image processing , image (mathematics) , enhanced data rates for gsm evolution , linguistics , philosophy , medicine , radiology
Rice is an agricultural commodity that is a staple food in Indonesia with hundreds of types of rice that have different characteristics. The type of rice can be distinguished from color and shape. The main feature that is dominant and can distinguish each type of rice is the color and shape. This feature is the main key in identifying types of rice. Identification is done by comparing the similarity of rice images using the value of color and shape features. The similarity can be determined through the difference in feature values between the query image and the database image. The closer the difference is to zero, the higher the level of similarity. The degree of similarity will affect the accuracy of image recognition at the time of identification. In this study, an analysis of the accuracy of image identification and measurement of computation time was carried out. Improved identification accuracy using the weighting of color and shape feature values. Extraction of the two value features using the invariant moment and color moment. Preprocessing before extraction using Grayscale, resize, edge Enhancement, Histogram Equalization. Clustering of rice image data using K-Means clustering. The results showed that the accuracy of identification with 400 rice image test data, reached more than 95% in the weighting scheme Ws (weighted Shape) = 40% and Wc (weighted color) = 60% with an average computing time of 5 milliseconds at 10 the cluster.