Determination of Sprout-Damaged Barley Using Thermal Imaging

Abstract

Pre-harvest sprouting is a major problem associated with cereal grains which results in lowering of end use quality. Pre-harvest sprouting affects the malting quality of barley.  The common methods to determine sprout damage are falling number, stirring number and amylograph peak viscosity, but these methods are time consuming.  There are other methods such as near infrared hyperspectral imaging and soft-x ray analysis which are still in the research stage.  Infrared thermal imaging technique to detect sprout damage is based on determining the changes in surface temperature distribution of grain which depends on the heat emission.  An infrared thermal camera was used in this study to determine whether sprout-damaged barley could be detected from healthy barley.  The results were analyzed using statistical and artificial neural network classifiers.  The classification accuracies were 78.7%, 78.9% and 88.5% for healthy; and 87.0%, 87.5% and 87% for sprouted kernels, using linear discriminant analysis, quadratic discriminant analysis and artificial neural network, respectively.  The results of the study show that thermal imaging has potential to determine sprout damage to barley.

Keywords: grain, barley, sprout-damaged, thermal imaging, classification, Canada