Determination of Effective and Specific Physical Features of Rice Varieties by Computer Vision In Exterior Quality Inspection

In this study, feature extraction processes were performed based on the image processing techniques using morphological, shape and color features for five different rice varieties of same brand. A total 75 thousand pieces grain obtained, including 15 each variety rice. Pre-processing operations applied to images made available extraction. 106 inferred from images; 12 morphological 4 obtained 90 spaces (RGB, HSV, L*a*b*, YCbCr, XYZ). addition, selected by ANOVA, X2 Gain Ratio tests useful determined. all tests, out features, 5 most effective specific roundness, compactness, factor 3, aspect ratio eccentricity. The listed in order following these features.