|Abstract in English:|
Qualitative characterization and cultivar identification is assuming a growing importance in agriculture. The assessment of these properties is important to deliver a product with specific characteristics that will satisfy industry and consumer expectations. The most used Technologies are based on visual evaluation made considering a reference model, therefore subjective, paired with more complex methods based on bi- or three-dimensional fruit shape. These have high costs and time of acquisition. This work tried to develop advanced and sophisticated instrumental methodologies, for online rapid application, for pomological
characterization and hazelnut cultivar classification. The research concerned four Italian cultivars namely: ‘Tonda Giffoni’, ‘Tonda Gentile Romana’, ‘Mortarella’ and ‘San Giovanni’. Specifically the following analyses were carried out: morphological analysis on fruit and seeds (weight, size), consistency test (max force) obtained by a digital dynamometer, spectro-colorimetric evaluation and shape analysis. For shape analysis, high resolution digital images of lateral view, polar view and random plane view for in-shell and kernel were obtained. The outline of each hazelnut was then processed with the elliptic Fourier analysis. The results showed that the cv. ‘San Giovanni’ presents a significantly harder shell compared with the other three cultivars (Analysis of Variance, ANOVA). The colorimetric test showed (.E > 6, perceivable by human eye) important differences between the cultivars ‘Tonda Giffoni’, ‘Tonda Gentile Romana’ and ‘Mortarella’. Moreover cultivar identification via shape analysis showed good results for correct classification (PLSDA) for the random plane view of in-shell (87.8%). In conclusion the studied
methodologies, even if applied in a pilot research, could be considered as a base for a broader evaluation taking into account the possibility of operating an efficient classification and characterization for a merceological and qualitative product selection.