A novel dataset and deep learning object detection benchmark for grapevine pest surveillance

Irrigation of vineyards is a matter of controversial arguments at areas of high quality wine production. Besides, the effects of the water in the plant are closer related to the water availability than to the irrigation regime.

Grapevine phenology has advanced across many regions, nationally and internationally, in recent decades under the influence of increasing temperatures, resulting in earlier
vintages (Jones and Davis, 2000, Petrie and Sadras, 2008, Tomasi et al., 2011, Webb et al., 2011. Earlier vintages have several ramifications for the wine industry. There are direct implications on quality, due to the fruit ripening during the hotter conditions of summer and early autumn, which then impacts grape composition and wine style (Sadras et al., 2013, Buttrose et al., 1971, Mira de Ordũna, 2010). There are also indirect implications where the fruit is perceived to ripen at a faster rate and the crop reach optimum maturity over a shorter period (Coulter et al., 2016).

Wine quality is defined by sensory and physico-chemical characteristics. In particular, sensory features are very important since they strongly condition wine acceptability by consumers. However, the evaluation of sensory quality can be subjective, unless performed by a tasting panel of experienced tasters. Therefore, it is of great relevance to establish relationships between objective chemical parameters and sensory perceptions, even though the complexity of wine composition makes it difficult. In this sense, more reliable relationships can be found for a particular wine typology or variety. The present study aimed to predict the perceived sensory quality from the physico-chemical parameters of ‘Barbera d’Asti’ DOCG red wines (Italy).

Wine grapes exposed to wildfire smoke have resulted in wines with burnt and ashy sensory characteristics¹, that are undesirable qualities in wine. In extreme wildfire events, this can lead to total loss of grape crop. Currently there are no effective solutions in the market to prevent the uptake of smoke compounds into grapes. In this study, previously developed innovative film coatings were tested to analyze their effectiveness in reducing smoke phenol absorption². Four different cellulose nanofiber-based film types were investigated.

During red wine ageing or conservation, color and taste change and astringency tends to reduce. These changes result from reactions of flavan-3-ols and/or anthocyanins among which condensation reactions with acetaldehyde are particularly important. The full characterization of these reactions has not been fully achieved because of difficulties in extracting and separating the newly formed compounds directly from wine. Model solutions mimicking food products constitute a simplified medium for their exploration, allowing the detection of the newly formed compounds, their isolation, and their structure elucidation.