This study aimed at assessing and classification of some grain varieties, on the basis of spectral data and indices calculated from aerial images. Were studied 18 varieties of cereals as follows: barley - Jallon, Tektoo, Saphira; wheat - Moison, Lukullus, Antonius, Linus, Atoupic, Illico, Sosthene, Avenue, G.K. Bekes, Akteur, Alex, Exotic, Messino, Patras; triticale - Trismart. The images were taken with the drone, model DJI Phantom series, in principal growth stage 3, and stem elongation, 30 – 31 BBCH code. By analyzing the images, they were obtained of primary spectral values in RGB system. Based on their values were calculated normalized rgb values, and also equivalent values were determined in HSB system. Based on primary spectral values (RGB), of the normalized values (rgb), and of the values in HSB system, specific indices were calculated for the characterization of vegetation: NDI (normalized difference index), INT (Intensity) and DGCI (dark green color index). ANOVA, revealed the source of variance and degree of statistical certainty of results (F>F crit; p<< 0.001). Through Cluster analysis (based on Euclidean distances) was performed classification of the varieties studied, based on the values of indices determined, and in relation to the degree of their values similarity. The 18 varieties studied were grouped into two clusters, with 5 in clusters and three independent positions, safely statistics (Cophenetic coefficient = 0.801).

aerial images, cereal grasses, classification, imaging, physiological indices

environmental engineering

Presentation: poster