Fractal geometry-based image analysis of grapevine leaves using the box counting algorithm

An image analysis method based on the box counting algorithm was evaluated for its potential to characterize grapevine leaves. Although vine leaves lack the self-similarity of the theoretical fractals, leaves are candidates for characterization using fractal analysis because oftheir highly complex structure. The results showed for 11 Sangiovese-related genotypes fractal dimensions significantly being different except for (1) Prugnolo acerbo, Prugnolo dolce and Prugnolo medio and (2) Chiantino, Brunelletto and Morellino di Scansano, which have shown a high similarity in agreement with previous studies. Fractal dimension calculated for leaves of Sangiovese R10 grown in very different environments did not show any statistically significant modification revealing that fractal dimension can be considered environment-independent. Consequently, fractal dimension could be used as a descriptive, scale-invariant, condensed, morphological parameter in ampelographic research. The role ofthe fractal dimension as an additional morphological parameter in future ampelographic classification schemes of grapevine leaves is discussed.