Algorithm development for grain kernel identification

Abstract

A digital image analysis algorithm was developed to facilitate classification of individual cereal grain kernels (barley, Canada Western Amber Durum (CWAD) wheat, Canada Western Red Spring (CWRS) wheat, oats, and rye). A total of 230 features (51 morphological, 123 color, and 56 textural) were extracted from 7500 high resolution color images of each type of grain using the developed algorithm. A four-layer back-propagation network (BPN) and k-nearest neighbor statistical classifier were evaluated for classification accuracies. The BPN used a sigmoid scaling function for input nodes and sigmoid activation function for nodes in the hidden layers. The data for statistical analysis was scaled using a normalizing function. Five different data sets were used for training, testing, and validation. The neural network based classifier outperformed the statistical classifier for all grain types. The average classification accuracies using BPN were 98.2, 90.9, 98.6, 98.4, and 99.0% for barley, CWAD wheat, CWRS wheat, oats, and rye, respectively. For the statistical classifier, the average classification accuracies were 85.1, 88.9, 96.9, 95.0, and 96.4% for barley, CWAD wheat, CWRS wheat, oats, and rye, respectively.