Inspection of Unpolished Rice by Using Image Sensing

Abstract

In this paper, we propose a new system of unpolished rice inspection by image sensing. The system classifies unpolished rice into 6 categories : speck rice, hull rice, ungrown rice, death rice, cracked rice and normal rice. To this end, the system employs the subspace method which uses five features : the gravity of the histogram, the variance of the histogram, the area of the object, the ratio of the major axis to the minor axis, and the area of the bright part. Some experimental results using a simple experimental system are also reported, which indicate effectiveness of the proposed method.