Discrimination of grapevine (Vitis vinifera L.) leaf shape by fractal spectrum.

A technique was developed (1) to digitise grapevine leaves, (2) to split the images obtained in the three components of the RGB (red, green and blue) colour system, obtain the fractal spectrum for each colour component of the leaf and (3) to calculate 15 different fractal parameters. The system, consisting of a scanning device, a personal computer and the code written in the C language, was then tested to characterize and identify 12 Sangiovese-related grapevine accessions. The results enabled us to distinguish between all accessions with a better discrimination than that obtained in previous studies with molecular markers or elliptic Fourier analysis. More, all the fractal parameters calculated for leaves of Sangiovese R10 grown in very different environments did not show any significant modification, revealing that fractal features can be considered environment-independent. The fractal analysis approach proposed, on the basis of the results obtained, gives the opportunity to verify the characters of distinction, uniformity and stability (DUS) requested by the Union for the Protection of Plant Varieties (UPOV) before plant breeders rights are granted.