Milling quality index for predicting flour yield

Abstract

The aim of the project was to devise a procedure, based on image analysis, for predicting the flour yield potential of a sample of wheat, which is superior to the existing unsatisfactory 'specific weight' measurement. Procedures were devised whereby grains could be cut in the plane of the crease so that a consistent face could be presented to the camera for TV. imaging. By this means measurements related to endosperm content could be made. At the same time as these measurements other shape factors were also recorded and relationships among them were derived. Milling yield was assessed on laboratory Buhler mills operated in a standard fashion and grade colour figures for all machine flours were obtained. On a series of ten wheats, chosen to represent a wide range of milling types, milling extraction rate correlated with one shape factor with a coefficient of 0.78. Inclusion of a second factor in a stepwise regression increased the correlation coefficient to 0.92.

A high correlation was also found with specific weight but the relationship was negative whereas a positive relationship is expected and frequently relied upon as a predictor. The unexpected negative correlation between specific weight and flour yield illustrates the difficulty in predicting yield in a wide range of wheat types. With this in mind the performance of the image analysis system is particularly promising. The present system of grain preparation is unsuitable for routine application but if the use of the present image analysis regime with a larger number of wheat types is equally successful a simplified method of presentation can be devised.