Improving the quality of automated VIS–grading of Scots pine seeds using fuzzy logic algorithm

Abstract

The automation of Scots pine seeds grading in the visible wavelength region – VIS-grading – is promising for conducting breeding and genetic experiments. This will reduce time costs and increase the accuracy of seed color classification compared to organoleptic techniques. When controlling VIS-grading, it is necessary to accurately detect and process the optical signal reflected from a single seed. The signal is based on the wavelength and amplitude of the optical beam. Earlier, using a spectrometer for Scots pine seeds from a natural stand of the Pavlovsky district of the Voronezh region, Russia, the boundaries of three spectrometric groups were established. In the real VIS-grading process, it is necessary to take into account the probabilistic deviations of random values of wavelengths and amplitudes. Therefore, on the basis of the Mamdani fuzzy logic theory developed an analyzing algorithm for controlling the VIS-grading quality. The algorithm consists of a sequence of logical terms that clearly define the specified VIS-grading seeds spectrometric parameters by a combination of wavelength and amplitude. The efficiency of Scots pine seeds VIS-grading using the algorithm is 98.9%.