By Andi Anderson
Researchers at the University of Illinois Urbana-Champaign have developed innovative near-infrared (NIR) spectroscopy models to analyze corn kernels and sorghum biomass efficiently.
These models, enhanced by machine learning, offer a faster, more accurate, and cost-effective alternative to traditional laboratory testing methods, which are often time-consuming, complex, and expensive.
In the first of two studies, the researchers created a global model to analyze moisture and protein content in corn kernels, crucial for the grain processing industry.
Using NIR spectroscopy, the team collected samples from seven countries—Argentina, Brazil, India, Indonesia, Serbia, Tunisia, and the USA—to develop a versatile model that adapts to different environmental conditions and agricultural practices.
The model combines gradient-boosting machines with partial least squares regression, providing reliable and accurate results across various locations.
The second study focused on sorghum biomass, a renewable and cost-effective feedstock for biofuel production. The researchers successfully used NIR spectroscopy to predict key features like moisture, ash, and lignin content.
This rapid and efficient method can significantly benefit biofuel, breeding, and other relevant industries by allowing quick and accurate biomass characterization.
According to the research team, NIR spectroscopy offers significant advantages over traditional lab analysis, including speed, environmental sustainability, and the ability to analyze multiple features without destroying samples.
This technology also provides flexibility, as samples can be scanned directly on the production line, making it highly practical for industrial use.
Photo Credit: gettyimages-nicexray
Categories: Illinois, Education