A computer vision approach for grade identification of rice bran

Inspection of food quality is an important operation in food and agro industries. Nowadays computer vision is frequently used for such operations as it can provide fast, economical, non-invasive, consistent and objective assessment. This paper presents a study on identifying the qualitative grades of rice bran using computer vision. The study is performed using three samples of rice bran collected from rice mills along with their test reports to confirm their qualitative difference. The images of individual samples were captured in a controlled illumination environment. The image features were extracted from the cropped images after the required color conversion. The constructed feature sets were subjected to principle component analysis (PCA) for observing the cluster formation and also the K-Means cluster analysis to derive the cluster centers. The clustering analysis results show the potential of the presented method for identification of rice bran grades.