Submitted to: Getreide Mehl Und Brot
Publication Type: Trade Journal
Publication Acceptance Date: January 1, 2002
Publication Date: January 1, 2002
Interpretive Summary: Near-infrared analysis of cereals has been practiced for more than 25 years. More recent has been its introduction as a tool for single kernel analysis. This additional information can contribute to the revamping of U.S. grain standards so that more emphasis is placed on end-use quality. Single kernel analysis by near-infrared spectroscopy offers the possibility to measure certain intrinsic properties, such as protein content, moisture content, wheat class and level of damage, in a rapid and non-destructive manner. Other applications include the identification and segregation of mold-infected or insect-infested kernels, and the identification and subsequent propagation of seeds with desirable traits in plant breeding programs. Advances in hardware and software in the past decade have resulted in scan rates of one kernel per second, with faster rates anticipated in the future. The present paper provides an overview of single kernel near-infrared research conducted at USDA-ARS facilities during the past 8 years and the future direction of this research. Beneficiaries to these projects include commercial wheat traders, millers, government inspectors, and seed companies.
Technical Abstract: As popularity in whole grain near-infrared (NIR) analyzers has increased in the past ten years, so too has an interest in using this technology at the level of a single kernel of grain. Research on the optical properties of single kernels of wheat has been underway in USDA laboratories at Beltsville, Maryland and Manhattan, Kansas. At Beltsville, single kernel research has consisted of wheat hardness, classification, and protein content by NIR transmittance, protein content and classification by NIR reflectance (1100-2500 nm), protein content of bulk samples by mathematical combination of single kernel protein predictions, and identification of scab-damaged kernels by hyperspectral image analysis. At Manhattan, single kernel NIR research has included refinement of the reflectance procedure to identify difficult-to-classify red and white kernels, prediction of protein content, detection of internal insect larvae infestation, detection of wheat scab, detection of vitreous kernels, and detection of heat damage, with all performed at near real-time conditions of the single kernel wheat characterization system (SKCS). These studies are discussed in context of their ramifications to official grading and classification, and to quality assessment through knowledge of kernel-to-kernel variability of intrinsic properties.