Title: Technical Note: Measuring Grain and Insect Characteristics using NIR Laser Array Technology Authors
|Praevium Research, Inc.|
Submitted to: Applied Engineering in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: September 1, 2009
Publication Date: January 1, 2010
Repository URL: http://www.ars.usda.gov/SP2UserFiles/Place/54300520/423.MeasuringgrainandinsectcharusingNIR.pdf
Citation: Dowell, F.E., Maghirang, E.B., Jayaraman, V. 2010. Technical Note: Measuring Grain and Insect Characteristics using NIR Laser Array Technology. Applied Engineering in Agriculture. 26(1):165-169. Interpretive Summary: A simple, low-cost, laser-based spectrometer was evaluated for measuring various grain and insect traits. This system is potentially much faster, cheaper, and more accurate than conventional near-infrared spectrometers. We evaluated the accuracy of this system for measuring wheat hardness, protein content, moisture content, and waxy character. We also used the system to determine the sex of tsetse fly pupae for potential sterile insect technique eradication programs. The laser cluster system predicted wheat hardness, moisture content, and fly sex with an accuracy similar to the near-infrared system, but predicted other traits with slightly less accuracy. The laser cluster system was limited to 8 wavelengths, and the accuracy of predicting other traits may be improved if different wavelengths were selected. This technology may provide a low-cost alternative for measuring some grain and insect traits.
Technical Abstract: The potential of using an eight-wavelength near-infrared (NIR) laser cluster spectrometer for measuring wheat quality (hardness index, protein content, moisture content, and waxy character) and determining tsetse fly pupae sex was investigated and compared to a commercial single kernel near infrared (SKNIR) system. Wheat hardness was predicted accurately by both NIR systems and results were in close agreement with reference values. Predicted protein content followed the same trend as the reference values, but the laser cluster system over-predicted low protein content values and under-predicted high values by about 1 percentage point. The accuracy of predicting moisture content by either system was similar with predicted values within 0.5% moisture content of the reference values. Waxy character was predicted by the laser system with less accuracy than the SKNIR system, but tsetse fly pupae sex was predicted with similar accuracies for both systems. Prediction equations derived from the laser spectra show that wavelengths influencing classification models generally agree with published literature. Thus, this research shows that a NIR laser cluster system can be used to predict some grain and insect traits with acceptable accuracy, and some predictions can likely be improved if other wavelengths are used in the laser cluster system.