Soybean sorting based on protein content using X-ray fluorescence spectrometry.

Area of Science:

• Agricultural Science
• Analytical Chemistry
• Spectroscopy

Background:

• Accurate determination of soybean protein content is crucial for agricultural and industrial applications.
• Traditional methods for protein analysis can be time-consuming and labor-intensive.
• Exploring rapid, non-destructive analytical techniques is essential for efficient crop assessment.

Purpose of the Study:

• To evaluate the performance of an energy-dispersive X-ray fluorescence (XRF) sensor for classifying soybean based on protein content.
• To investigate the utility of sulfur (S) signals and other XRF spectral features as proxies for inferring soybean protein levels.

Main Methods:

• Optimization of sample preparation and equipment settings for enhanced detection of sulfur and other relevant XRF emission lines.
• Development of a logistic regression model utilizing XRF spectra and established protein content data for classification.
• Validation of the developed model using an independent set of soybean samples.

Main Results:

• The logistic regression model achieved global accuracies of 0.83 for the training set and 0.81 for the test set.
• Corresponding kappa indices were 0.66 (training) and 0.61 (test), indicating satisfactory classification performance.
• XRF spectral features, particularly sulfur signals, demonstrated effectiveness in inferring soybean protein content.

Conclusions:

• Energy-dispersive X-ray fluorescence (XRF) shows satisfactory performance for classifying soybean protein content.
• XRF spectral features, including sulfur signals, can be reliably applied for screening protein levels in soybeans.
• This technique offers a promising avenue for rapid and efficient protein analysis in soybean breeding and quality control.