Non-destructive Estimation of Maize (Zea mays L.) Kernel Hardness by Means of an X-ray Micro-computed Tomography (μCT) Density Calibration

Abstract

An X-ray micro-computed tomography (μCT) density calibration was constructed for whole maize (Zea mays L.) kernels from polymers ranging in absolute densities from 0.9 to 2.2 g cm−3. The resulting linear equation was used to estimate the densities of two regions-of-interest, i.e. vitreous and floury endosperm, as well as that of the entire maize kernel. The sample set comprised 16 maize kernels (eight hard and eight soft). Validation of the entire kernel density was performed by comparing estimated and measured (actual) masses (r = 0.99; standard error of measurement = 0.01 g). In addition, percentage cavity (%cavity) and percentage porosity (%porosity) were quantified from the X-ray images. As determined with analysis of variance, floury, vitreous and entire kernel endosperm densities as well as %cavity and %porosity significantly (P < 0.05) contributed to the variation within the hardness classes. Hardness classification was attempted using a receiver operating characteristic curve. Threshold values of 1.48, 1.67 and 1.30 g cm−3 were determined for the entire kernel, vitreous and floury endosperm densities, respectively, at a maximum of 100 % sensitivity and specificity. Classification was possible from %porosity values of the entire kernel, which are easier to determine, at 88 % accuracy. Efficient maize kernel hardness classification is required by the dry-milling industry when maize is milled into a meal and used as a food source. Optimum quality and yield can only be obtained during the milling process if maize of appropriate hardness is used as raw material.