Sorghum Grains Grading for Food, Feed, and Fuel Using NIR Spectroscopy

Area of Science:

• Agricultural Science
• Analytical Chemistry
• Biochemistry

Background:

• Near-infrared (NIR) spectroscopy offers a rapid, non-destructive method for cereal grain quality assessment.
• Limited investigation exists on NIR's feasibility for critical biochemicals crucial for food, feed, and fuel classifications.

Purpose of the Study:

• To investigate the feasibility of Fourier-transform (FT) NIR spectroscopy for determining eight key biochemicals in sorghum.
• To evaluate the impact of sample type (hulled/hull-less grain flours, whole grains) on prediction accuracy.

Main Methods:

• Utilized FT-NIR spectroscopy on four sorghum sample types from 20 hybrids.
• Employed partial least squares regression (PLSR) to build prediction models using spectral and wet-chemical data.

Main Results:

• Sorghum grain morphology and sample format significantly influenced biochemical prediction accuracy.
• Grain flour samples generally yielded improved predictions compared to whole grains.
• Hull-less grain flours enhanced predictions for tannin, cellulose, and hemicellulose over hulled flours.

Conclusions:

• FT-NIR spectroscopy is a viable tool for non-destructive sorghum biochemical analysis.
• Sample preparation (flour vs. whole grain) impacts prediction performance, with flours often superior.
• Hull-less grain flours offer advantages for specific biochemical predictions, aiding diverse sorghum applications.