Image processing techniques for grading & classification of rice

A relatively faster computer vision system has been discussed to analyze and sort rice kernels. A series of measurements were done using image processing techniques on three varieties of Indian rice namely Markfed Supreme, Markfed Golden (export quality), Hafed Basmati. Area, perimeter, maximum length, maximum width, compactness and elongation were measured. Further, separating the rice varieties by their shape difference was examined. The computer vision system developed has been able to sort rice into sound, cracked, chalky, broken and damaged kernels with an accuracy ranging from 90-95%.