Identification of varieties of rice based on sparse representation

An identification method based on sparse representation was proposed for discriminating the varieties of rice precisely. The rice images of six varieties such as long glutinous rice, round glutinous rice, non-glutinous rice, Thailand aromatic rice, red aromatic rice and black rice were taken as the research objects. To represent single rice kernel, its color and morphological characters were extracted. For each varieties, 50 grains of rice were selected randomly as the training samples, and 200 grains of rice were treated as the testing samples. All of the training samples made up the data dictionary of the sparse representation, and the projection of the testing sample on the data dictionary was calculated. The breed, which had the minimum projection error, would be regarded as the right kind of rice. At last, the identifying results on the proposed method were analyzed and compared with those of the BP network and SVM. Experimental results demonstrated that the overall identification accuracy of the proposed algorithm for the six rice breeds was 99.6%, which was the best classification effect among three methods. Therefore, the proposed method can provide a new effective method for identification of rice breed.