Classification of Milled Rice Using Image Processing

Classification of different types of rice is carried out in this study using metaheuristic classification approaches.13 different rice samples are considered. Images of milled rice are acquired using a computer vision system. Feature Extraction methods are used to extract fifty seven features including five shape and size features, forty eight color features and four texture features from color images of individual rice samples. Four different metaheuristic classification techniques including Artificial Neural Network, Support Vector Machine, Decision Tree and Bayesian Network are utilized to classify milled rice samples. Results indicated that Artificial Neural Network had the highest classification accuracy (92.307 %) followed by Support Vector Machine (90.384 %), Bayesian Network (82.692 %) and Decision Tree (59.615 %), had the higher accuracy, respectively.