Image analysis of broken rice grains of Khao Dawk Mali rice

Abstract

In process of rice grain milling, rice grains are sorted by size into many categories for sale at different prices. By observation on “small broken”, which is a low-quality category of sorted result, we found that it composes of significant amounts of more expensive rice grains. In this research, we propose a method to evaluate broken rice grains in order to make higher profit from its higher quality portion by image analysis. Our algorithm is to categorize “small broken” into four types: small broken, broken, big broken and head rice, which are classes described by the Department of Rice, Thailand. Least-Square Support Vector Machine (LS-SVM) with Radius Basis Function (RBF) kernel is used as a classifier in the algorithm. The accuracy of the algorithm is 98.20%.