Search

It is crucial for winegrowers to make informed decisions about the optimum time to harvest the grapes to ensure the production of premium wines. Global warming contributes to decreasing acidity and increasing sugar levels in grapes, resulting in bland wines with high contents of alcohol. Predicting quality in viticulture is thus pivotal. To assess the average ripeness, typically a sample of one hundred berries representative for the entire vineyard is collected. However, this process, along with the subsequent detailed must analysis, is time consuming and expensive. This study focusses on predicting essential quality parameters like sugar and acid content in Vitis vinifera (L.) varieties 'Chardonnay', 'Riesling', 'Dornfelder', and 'Pinot Noir'. A small near-infrared spectrometer was used measuring non-destructively in the wavelength range from 1 100 nm to 1 350 nm while the reference contents were measured using high-performance liquid chromatography. Chemometric models were developed employing partial least squares regression and using spectra of all four grapevine varieties, spectra gained from berries of the same colour, or from the individual varieties. The models exhibited high accuracy in predicting main quality-determining parameters in independent test sets. On average, the model regression coefficients exceeded 93% for the sugars fructose and glucose, 86% for malic acid, and 73% for tartaric acid. Using these models, prediction accuracies revealed the ability to forecast individual sugar contents within an range of ± 6.97 g/L to ± 10.08 g/L, and malic acid within ± 2.01 g/L to ± 3.69 g/L. This approach indicates the potential to develop robust models by incorporating spectra from diverse grape varieties and berries of different colours. Such insight is crucial for the potential widespread adoption of a handheld near-infrared sensor, possibly integrated into devices used in everyday life, like smartphones. A server-side and cloud-based solution for pre-processing and modelling could thus avoid pitfalls of using near-infrared sensors on unknown varieties and in diverse wine-producing regions. [ABSTRACT FROM AUTHOR]

Balanced and stable yield is a major trait in grapevine breeding and breeding research. Grapevine yield hereby is a complex quantitative trait, as it is influenced by multiple plant parameters, like berry size, number of berries per bunch, number of bunches per shoot, management, and environmental factors. In the current breeding process, the complexity of this trait has shown that a classification according to descriptive factors for marker development is only possible to a limited extent. Precise field phenotyping of yield-related traits is the basic prerequisite to be able to measure such quantitative traits. This, however, is the major bottleneck due to labor, time and constrains of plant material in the breeding process. For this reason, one of our main goals with the newly developed phenotyping platform PHENOquad with its multisensor system PHENOboxx is to improve phenotyping efficiency of grapevine yield to overcome the phenotyping bottleneck. The newly developed embedded vision system PHENOboxx is mounted on an "all-terrain vehicle (ATV)". This allows a fast data acquisition on a large number of individual vines. In order to evaluate the yield potential of breeding material in comparison to established grapevine cultivars, various yield-related parameters of the vines are quantified directly in the field with high spatial and temporal resolution. As key parameters for yield-related phenotyping, the number of shoots, bunches, berries and the weight of dormant pruning wood was identified. The image data acquired are annotated to train the artificial intelligence (AI). Within the process, the image analysis results are compared to annotated ground truth data and correlated with the field reference data. We expect to increase the precision, target specificity and throughput of screening grapevine material without reducing its accuracy over time by using the PHENOquad. In addition, a weighting of yield-relevant parameters would be possible. This opens up new possibilities for efficient plant evaluation in the scope of grapevine breeding. Also new application possibilities for precision viticulture are conceivable. [ABSTRACT FROM AUTHOR]

Grapevine is a perennial woody plant whose characteristics, in addition to marker-assisted selection (MAS) on resistances, are mainly recorded in the field during the breeding steps. For material selection in breeding long-term observation and recording of the overall vitality, yield, phenology, growth architecture, resistance against diseases, and quality is done. However, the time window for recording characteristics during the growing season is narrow and the evaluation capacity is limited due to manual assessment. Thus, phenotypic characteristics cannot be documented in detail for early breeding stages for the large number of breeding material, which leads to a lack of information about the selected genotypes and later no retrospective data analysis of individual genotypes is possible. Therefore, a novel high throughput determination system is needed to increase the efficiency of phenotyping and to optimize data management within the breeding program. In the present study, the expanded phenotyping platform "Phenoliner 2.0" and the steps to develop such a phenotypic pipeline for breeding are shown. The platform, which is equipped with a prism-based simultaneous multispectral camera system consisting of a visible color channel and two near-infrared channels, was developed and tested in the vineyard to acquire images and spectra of grapevines. Compared to the previous version, the speed of data acquisition has been improved from one kmh to four kmh. The assignment of the sensor data is based on GPS information of the vine locations, which are recorded by a real-time kinematic GPS receiver mounted on the phenoliner. With the help of a plant detection algorithm, the absence of vines can be recognized and documented automatically. In order to develop the image analysis pipeline for the main phenotypic breeding traits: (1) vitality, (2) yield and (3) wood maturity, the system needs to be trained to distinguish among different structures of the grapevine (leaf, grape bunch, and stem) and between healthy and diseased structures. For this purpose, the recorded images were segmented and manually annotated. So far, the annotation software developed in this study is able to automatically recoginze leaf and grape bunch structures in the images and videos. To improve data management, the images and their results could be stored and linked to individual genotype metadata, such as breeding number, planting years, and crossing parents. The system can be applied to map phenotypic parameters of an entire vineyard. The information on vitality, yield, ripening, and resistances is important for further breeding steps and viticulture. [ABSTRACT FROM AUTHOR]

The most useful quality-determining factors for grape must are sugars and acids due to their importance for the production of high quality wine. Moreover, German wine law divides different categories according to the sugar content of the must depending on the EU defined wine-growing zones. Therefore, in most cooperatives the sugar content in particular is one basic factor for payment. Winegrowers have the possibility to manually measure the sugar and acid contents in their vineyards prior to harvest. However, these measurements are individual observations and do not represent the whole plot. Cooperatives check the quality of berries upon delivery for payment and further cellar logistics. An early detection of must characteristics prior to delivery on a grape harvester for example would improve the harvest and cellar logistics in big cooperatives. While there are already many applications for near-infrared spectroscopy to detect ingredients in biological samples, there is still no applicable detector for determining ripeness in viticulture. To close this gap, an apparatus was built and tested in the laboratory. It consists of a near-infrared sensor and a flow cell, recording spectra in the range of 1100 nm to 1350 nm of grape must. This must was collected from 100 berries from each of four different red and white Vitis vinifera (L.) varieties from different vineyards. The spectra were acquired throughout the growing season in 2021 in Rhineland-Palatinate (Germany). Reference values for the main sugars (glucose, fructose), as well as acids (malic-, tartaric acid) were determined with high-performance liquid chromatography. Using chemometric methods, i.e. partial least square regression and leave-one-out cross-validation, it was possible to estimate root mean square prediction errors (RMSEP) of 4.96 g/l and 0.60 g/l and a determination coefficient (R²) of 83% for sugars and acids, respectively. Coefficient of determination and RMSEPs were higher for red varieties compared to white varieties for both substance classes. This high prediction accuracy could enable automatic ripeness measurements during the harvesting process on the harvester, thus reducing the workload during harvest and offering more planning security. [ABSTRACT FROM AUTHOR]

With the rising trend of consumers being offered by start-up companies portable devices and applications for checking quality of purchased products, it appears of paramount importance to assess the reliability of miniaturized sensors embedded in such devices. Here, eight sensors were assessed for food fraud applications in skimmed milk powder. The performance was evaluated with dry- and wet-blended powders mimicking adulterated materials by addition of either ammonium sulfate, semicarbazide, or cornstarch in the range 0.5–10% of profit. The quality of the spectra was assessed for an adequate identification of the outliers prior to a deep assessment of performance for both non-targeted (soft independent modelling of class analogy, SIMCA) and targeted analyses (partial least square regression with orthogonal signal correction, OPLS). Here, we show that the sensors have generally difficulties in detecting adulterants at ca. 5% supplementation, and often fail in achieving adequate specificity and detection capability. This is a concern as they may mislead future users, particularly consumers, if they are intended to be developed for handheld devices available publicly in smartphone-based applications. [ABSTRACT FROM AUTHOR]