Identification of soybean seed varieties by spectral luminescence method

Optical spectral photoluminescent diagnostic methods, characterized by high speed and low implementation costs, can serve as an alternative to existing methods of varietal seed identification. The study aims to develop a methodology for identifying soybean seed varieties using a spectral luminescence method. The author used a diffraction spectrofluorimetric complex to measure the spectral excitation characteristics of photoluminescence in six soybean varieties: ultra-early maturing (Puma and Sayana), early maturing (Greya and Selena), and mid-maturing (Vilana and Vilana beta). Based on the obtained dependencies, he calculated integral and statistical parameters: integral absorptive capacity, mathematical expectation, variance, skewness, and kurtosis. The photoluminescence excitation characteristics of soybean seeds were found in the range between 330 and 475 nm, with primary peaks at 392 and 420 nm. For Gaussian decomposition, the spectra were converted to the frequency domain using the inverse Fourier transform. Based the research findings, the author selected the integral absorptive capacity in the range between 350 and 396 nm as the variety identification parameter. Skewness was recognized as the most effective statistical parameter for the unambiguous identification of three soybean varieties. Gaussian decomposition parameters – variance, skewness, and frequency-domain mathematical expectation – can be applied for the unique identification of all studied varieties. The methodology for identifying soybean seed varieties includes sample preparation, irradiation in the range between 330 and 475 nm, and registration of photoluminescent signals by a radiation receiver to obtain the U(λ) dependency and, based on it, calculate the varietal identification parameters: Gaussian amplitude and variance, frequency-domain mathematical expectation, integral absorptive capacity, and skewness. Narrow-spectrum LEDs and a broad-spectrum photodiode are used to implement the developed methodology. The characteristics of the U(λ) photo-signals help determine the varietal affiliation of the seeds. The results obtained can be used to develop a portable, high-speed variety control device for both laboratory and field conditions.