Fractal analysis for soybean leaves characterization.

The present study used fractal analysis for soybean leaves characterization (Glycine max (L.) Merr), Caro and Onix genotypes. The box-counting method was used to find the fractal dimensions of leaf shape in the two genotypes. The fractal dimensions were D = 1.627 at the Caro genotype, and respectively D = 1.509 at the Onix genotype. In the case of Caro genotype, correlation analysis showed average correlations between fractal dimension (D) and leaf width (W), r = 0.728, respectively between D and scan leaf area (SLA), r = 0.778 and low intensity correlations between D and perimeter of the leaves (Per), r = 0.624, respectively between D and the leaf length (L), r = 0.484. In the case of Onix genotype, high correlations were found between D and SLA (r = 0.862), between D and W (r = 0.812); average correlations between D and Per (r = 0.766), and respectively low correlations between D and L (r = 0.685). Based on D values, SLA prediction was possible under conditions of R² = 0.631, p << 0.001 for Caro genotype, and R² = 0.812, p << 0.001 respectively for Onix genotype. [ABSTRACT FROM AUTHOR]