A New Germinative Classification Model of Barley for Prediction of Malt Quality Amplified by a Near Infrared Transmission Spectroscopy Calibration for Vigour 'On Line' Both Implemented by Multivariate Data Analysis

摘要

A new germinative two-dimensional classification plot fully compatible to the current EBC analyses (EBC methods 3.5-3.7) is proposed for malting barley based on separate estimates for "vigour" (24 h germination) as abscissa with limits at 70% and 30% and for "viability" (72 h germination) as ordinate with limits at 98% and 92%. Early detection of germination by image analysis was improved by utilising the auto fluorescence of the root cap. The seven hierarchical germinative classes visualise the quality differences in a consistent way, ordering classes according to falling extract % and increasing wort β-glucan (mg/L). It was surprising to discover that significant barley Near Infrared Transmission (NIT) spectroscopy based Partial Least Squares Regression prediction models for "vigour" and "viability" were obtained after removing the PLSR outliers. The majority of these were found to be low in vigour. It was concluded after experimental validation that the physical-chemical structure of the seed, reflected by the correlation of the barley NIT spectral fingerprints to germination speed, is connected to the availability of substrate for germ growth. This is another aspect of the speed of malt modification. An automated combination instrument for measuring physical-chemical and seed germination parameters is suggested for quality control and to establish an on-line NIT calibration network for integrated germinative and malting quality classification.