Influence of growing location on features extracted from colour images of wheat and detection of foreign material represented by barley

Several studies have demonstrated potential application of machine vision systems in the grain industry for quality inspection, grading, and classification. However, the influence of colour image features from grain samples from different growing locations on the training and classification performance has not been investigated thoroughly. In this study, morphological (51), colour (123), and textural (56) features from colour images of Canada Western Red Spring (CWRS) wheat kernels were extracted to study the influence of growing location. Top features selected by STEPDISC procedure were used in statistical comparison. Most of the image features from different growing locations had significant differences; however, these differences did not affect the grain classification performances as tested by detecting foreign material, represented by barley.