Predicting nutritional composition of corn grain using near infrared reflectance spectroscopy

摘要

Corn grain samples collected over severalyears were used to develop equations for drymatter, crude protein, crude fiber, and etherextract (crude fat). Two computer models wereused to select samples having a range of spectra(presumably a range of nutrient values) fordeveloping calibration equations. Both methodsselected an unexpectedly small number ofsamples; however, only the prediction of crudefiber appears questionable. Althoughcoefficients of determination values are expectedto be low with a small number of samples,standard errors of validation and prediction areconsistently lower than those of standard(AOAC) methods, suggesting that thetechnology is reliable for nutrient analysis ofcorn grain.