Classification of bulk samples of cereal grains using machine vision

Digital image analysis algorithms were developed to classify bulk samples of Canada Western Red Spring (CWRS) wheat, Canada Western Amber Durum (CWAD) wheat, barley, oats, and rye using textural and colour features. The textural features of bulk samples were extracted from different colours, i.e., red R, green G, or blue B, and colour band combinations [black & white {(R + G + B)/3}; (3R + 2G + 1B)/6; (2R + 1G + 3B)/6; or (1R + 3G + 2B)/6] of images to determine the colour or colour band combination that gave the highest classification accuracies in cereal grains. The textural features extracted from the red colour band at maximum gray-level value 32 gave the highest classification accuracies in cereal grains (mean accuracy, the average of the classification accuracies of the above-mentioned cereal grains, was 100% when tested on an independent data set that had 10 500 grain kernels). When the original bulk images were partitioned into sub-images and textural or colour features extracted from the sub-images were used, the classification accuracies of cereal grains decreased compared to when the original bulk images were used. The mean accuracy was 100% when colour features of bulk samples were used for classification of cereal grains in an independent data set. © 1999 Silsoe Research Institute.