Seed morpho-colorimetric analysis by computer vision: a helpful tool to identify grapevine (Vitis vinifera L.) cultivars

Background and Aims: The Sardinian wine heritage consists of 151 cultivars; however, because of the different dialects within the same region, many of these cultivars can be considered as false attributions (synonyms/homonyms). The aim of this study was to use seed morpho-colorimetric features to discriminate among the grapevine cultivars and identify synonymy groups. Methods and Results: Over two seasons, 230 grapevine accessions belonging to 115 Sardinian native cultivars (54 black berries and 61 white berries) were collected. Seed images were acquired and analysed. Thirty-three morphocolorimetric attributes and 80 elliptic Fourier descriptors were assessed. The data were analysed by applying linear discriminant analysis to implement statistical classifiers able to identify the landraces. Fourteen black berry synonymy groups and eight independent cultivars, and 15 white berry synonymy groups and nine independent cultivars, were identified. Conclusions: These results confirmed the synonymy groups previously proposed and improved some of them with other local cultivars. Significance of the Study: Despite different historical and cultural consumption purposes, some cultivars belong to the same synonymy groups, causing apparent incongruence. Nevertheless, the results are consistent with the geographical distribution of the cultivars and with the historical and cultural knowledge in Sardinia.