Together with ensuring a stable yield, improving grape composition and aroma is the main goal of wine grape production management as it determines consumer acceptance and ultimately revenue. Understanding the triggers of the synthesis of aromatic components and finding methods to map their variability in the field can aid management practices during the season and planning selective harvest in views of maximizing benefit. Vegetation indices have been shown to track grape colour, sugar and acidity content but it has been demonstrated that aromatic components are the main drivers of the final palate of wine and are not correlated to sugar concentration. Leaf pigments such as chlorophyll, carotenoids and anthocyanins are involved in the metabolic pathways of aroma compounds in grapes. The physiological connections between grape aromatic components and primary and secondary photosynthetic pigments suggest that they could be used to detect processes related to aroma composition.

This study investigates the links between grape quality parameters such as aromatic components and image-quantified spectral indices and photosynthetic plant traits derived by physical model inversion methods. Two sets of high-spatial resolution hyperspectral and thermal imagery were collected with an unmanned platform at veraison and harvest. The variability found in the field was partly but not fully explained by the thermal-based crop water stress index as an indicator of water stress (r²= 0.51–0.58, p-value<0.01). Fluspect-CX leaf model was coupled to 4SAIL canopy model and inverted to map the main photosynthetic pigment groups and the fraction of pigments acting in photoprotection. Results obtained through radiative transfer model inversion outperformed traditional vegetation indices related to pigment content and degradation. We found statistically significant relationships between image-retrieved pigments and terpenoids responsible for wine aroma (p-value<0.005).

在确保稳定的产量的同时，改善葡萄的成分和香气是酿酒葡萄生产管理的主要目标，因为它决定了消费者的接受程度并最终决定了收入。了解芳族成分合成的诱因并找到方法来绘制其在田间的变异性，可以在本季节进行管理实践，并从最大化收益的角度规划选择性收获。植被指数已显示出可跟踪葡萄的颜色，糖和酸度的含量，但已证明芳香成分是葡萄酒最终口感的主要驱动力，并且与糖浓度无关。叶色素（例如叶绿素，类胡萝卜素和花色苷）参与葡萄中香气化合物的代谢途径。

本研究调查了通过物理模型反演方法得出的葡萄质量参数（如芳香成分和图像量化的光谱指数）与光合植物性状之间的联系。在无人值守的平台上收集和采集了两组高空间分辨率的高光谱和热成像图像。田间发现的变异性部分但没有完全解释为基于热的作物水分胁迫指数作为水分胁迫的指标（r ²= 0.51-0.58，p值<0.01）。将Fluspect-CX叶片模型与4SAIL冠层模型耦合并反转，以绘制主要光合色素基团和参与光保护作用的色素比例。通过辐射转移模型反演获得的结果优于与色素含量和降解有关的传统植被指数。我们发现图像提取的色素和负责葡萄酒香气的萜类化合物之间存在统计学上的显着关系（p值<0.005）。