The article is devoted to the study of the opportunities of evaluating the vitreousity of malted barley using machine vision and image processing. With the help of the developed hardware and software complex, experimental studies were conducted on samples of three different barley varieties. As a result, the optimal shooting mode was selected and an algorithm for processing digital images of barley grains was developed to determine the number of vitreous grains in the batch. In addition to classifying grains according to their vitreousity, the proposed approach also allows to evaluate the sample's uniformity and, thus, to identify higher-quality barley. As a result of additional research, it was found that the grain orientation adds an error of no more than 5%. High repeatability of results and the accuracy of the algorithm are characterized by variation coefficient which is 1.1%.
|