A multispectral sorting device for isolating single wheat kernels with high protein content

Authors: Haff, Ronald - Ron; Pearson, Thomas; Maghirang, Elizabeth

Submitted to: Journal of Food Measurement and Characterization
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/19/2013
Publication Date: 12/2/2013
Citation: Haff, R.P., Pearson, T.C., Maghirang, E.B. 2013. A multispectral sorting device for isolating single wheat kernels with high protein content. Journal of Food Measurement and Characterization. 7(4):149-157.

Interpretive Summary: Automatic sorting of single wheat kernels by protein content was demonstrated using two novel multispectral sorting devices, one measuring visible light wavelengths and the other measuring wavelengths in the near infrared (NIR). The data were acquired by rapidly (~12 kHz) and sequentially pulsing LEDs at nine wavelengths and measuring the amount of light reflected from the wheat kernels at each wavelength as they drop off a feeder chute. A microcontroller was used to control the sorter electronics, perform signal processing, and apply classifications. Sorting was accomplished through activation of an air valve. The instruments were tested by sorting 13 hard red winter samples to select high protein kernels. The resulting sample of high protein kernels was sorted a second and third time to further increase average sample protein content. For the visible light instrument, the sorted samples had 0.99%, 1.60%, and 1.86% higher protein than the original samples after each sort, respectively. Similarly, after sorting with the NIR based LED instrument, sorted samples had 0.60%, 1.25%, and 1.68% higher protein than the original samples after one, two, and three sorts, respectively. The simple LED-based sorter is very economical as compared to most other small scale sorting systems, with material costs for the prototype of around $6000. This sorter would likely be most effectively employed to select desired traits of small lots of seed (e.g., breeder size samples), assist in sample purification, and help breeders in selecting the kernels with higher protein.

Technical Abstract: Automated sorting of single wheat kernels according to protein content was demonstrated using two novel multispectral sorting devices with different spectral ranges; 470-1070 nm (silicone based detector) and 910nm-1550 nm (InGaAs based detector). The multispectral data were acquired by rapidly (~12 kHz) and sequentially pulsing LEDs at nine wavelengths and measuring the amount of light reflected from the wheat kernels at each wavelength as they drop off a feeder chute. A microcontroller was used to direct the LED pulses, digitize the analog signal from the photodiode, perform signal processing, and apply classifications. Sorting was accomplished through activation of an air valve. The instruments were tested by sorting 13 hard red winter samples to select high protein kernels. The resulting sample of high protein kernels was sorted a second and third time to further increase average sample protein content. For the silicone based instrument, the sorted samples had 0.99%, 1.60%, and 1.86% higher protein than the original samples after each sort, respectively. Similarly, after sorting with the InGaAs based LED instrument, sorted samples had 0.60%, 1.25%, and 1.68% higher protein than the original samples after one, two, and three sorts, respectively. The simple LED-based sorter is very economical as compared to most other small scale sorting systems, with material costs for the prototype of around $6000. This sorter would likely be most effectively employed to select desired traits of small lots of seed (e.g., breeder size samples), assist in sample purification, and help breeders in selecting the kernels with higher protein.