Application of image texture analysis for varietal classification of barley

Abstract

This paper presents the results of a study into the use of the texture parameters of barley kernel images in varietal classification. A total of more than 270 textures have been calculated from the surface of single kernels and bulk grain. The measurements were performed in four channels from a 24 bit image. The results were processed statistically by variable reduction and general discriminant analysis. Classification accuracy was more than 99%. © 2012 Institute of Agrophysics, Polish Academy of Sciences.