А. И. Новиков In adaptive farming, focused on stress tolerance and resource efficiency, pre-sowing seed quality assessment of shelterbelt crops is critical. The study evaluates the predictive potential of normalized spectral indices of the seed coat (epidermis) in the RGB space for forecasting laboratory-container germination. The research uses as an example a breeding form of Scots pine (Pinus sylvestris L. cv. Negorelskaya), which is used in protective afforestation. Seven normalized indices minimizing the influence of overall brightness were analyzed: [(R-G-B)/(R+G)]², [(R-B)/(R+B)]², [(G-B)/(G+B)]², [(G-R)/(G+R)]², [(R+G-B)/(R+G+B)]², [(R+B-G)/(R+G+B)]², [(G+B-R)/(R+G+B)]². Scanner images were obtained for more than 1 000 individual seeds, and container germination was assessed on the 50th day. Statistically significant differences (p<0.0001-0.0165) in the distributions of all indices were established between zero- and non-zero-germination groups. Stable spectral patterns associated with low germination potential were identified: increased relative reflectance in the red and green channels compared to the blue channel and decreased reflectance in the blue channel relative to the red and green ones. The findings demonstrate the possibility of integrating simple, low-cost spectral markers based on RGB scanning into seed technology passports. This approach enables non-destructive selection of viable seed material with predictable properties. It is a key element of the “Target Plant” concept for establishing adaptive, stress-tolerant forest shelterbelts.
Научная статья 
Pinus sylvestris L.