Hardness, of wheat grains, is one of the most important quality characteristics used in wheat classification and determination of its marketing value. So, the key objective of this investigation applies a non-destructive method like infrared technique as an alternative method of destructive methods to assess hardness of wheat grains. The hardness characteristic was measured by two destructive methods Single-Kernel Characterization System (SKCS) and Instron Universal Testing Machine (IUTM), as reference values. Infrared technique was used to develop NIR calibration and validation model using the partial least squares (PLS) regression to assess wheat grain hardness. The best calibration and validation model for assess hardness of wheat grains were observed throughout the reference method Instron Universal Testing Machine (IUTM) within the wavelength range 950 to 1650 nm with 6 principal components (PCs) and pretreatment by Savitzky-Golay second derivative (S.G. 2supnd/sup). Where, the optimum PLS was recorded at the lowest standard error of prediction (SEP) 3.92 N with the maximum value of coefficient of prediction (Rsup2P/sup ≈ 0.91) and sufficient value of the relative prediction deviation (RPD ≈ 3.35). The accuracy of the prediction model was sufficient to use NIRS technique as a nondestructive method to estimate hardness of wheat grains for different varieties of the wheat.
展开▼
