terclim by ICS banner
IVES 9 IVES Conference Series 9 A novel dataset and deep learning object detection benchmark for grapevine pest surveillance

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

Abstract

Flavescence dorée (FD) stands out as a significant grapevine disease with severe implications for vineyards. The American grapevine leafhopper (Scaphoideus titanus) serves as the primary vector, transmitting the pathogen that causes yield losses and elevated costs linked to uprooting and replanting. Another potential vector of FD is the mosaic leafhopper, Orientus ishidae, commonly found in agroecosystems. The current monitoring approach involves periodic human identification of chromotropic traps, a labor-intensive and time-consuming process.

Therefore, there is a compelling need to develop an automatic pest detection system, leveraging the recent progress in computer vision and deep learning techniques. However, the current progress in developing such a system is hindered by the lack of effective datasets to serve as ground-truth data for the training process.

To fill this gap, our study contributes a fully annotated dataset of S. titanus and Or. ishidae from yellow sticky traps. The dataset comprises more than 400 images, with 1000 identification per class. Guided by entomologists, the annotation task involved defining bounding boxes around relevant insects with corresponding class labels.

We trained and compared the performance of state-of-the-art object detection algorithms (YOLOv8 and Faster R-CNN). Pre-processing included automatic cropping to eliminate irrelevant background information and image enhancements to improve overall quality. Additionally, we tested the impact of altering image resolution, data augmentation, and single-class detection. Preliminary results achieved a high detection accuracy, with mAP@50 and F1-score above 90%, and mAP@50-95 around 70%, allowing a first deployment as an automatic annotation support tool.

DOI:

Publication date: June 14, 2024

Issue: Open GPB 2024

Type: Poster

Authors

Giorgio Checola1*, Paolo Sonego1, Valerio Mazzoni2, Franca Ghidoni3, Alberto Gelmetti3, Pietro Franceschi1

1 Research and Innovation Centre, Digital Agriculture Unit, Fondazione Edmund Mach, S. Michele all’Adige, TN, Italy
2 Research and Innovation Centre, Plant Protection Unit, Fondazione Edmund Mach, S. Michele all’Adige, TN, Italy
3 Technology Transfer Centre, Viticulture Unit, Fondazione Edmund Mach, S. Michele all’Adige, TN, Italy

Contact the author*

Keywords

insect detection, deep learning, smart pest monitoring, flavescence dorée, insect traps

Tags

IVES Conference Series | Open GPB | Open GPB 2024

Citation

Related articles…

Étude des relations sol-vigne sur le vignoble de Côte Rôtie

La topographie du vignoble de Côte Rôtie, la prédominance de la non culture ainsi que la structure très légère des sols amènent les vignerons à s’interroger sur l’entretien du sol, la conduite de la fertilisation de leurs parcelles ainsi que sur le développement racinaire de la vigne.

Study on the impact of clone on the varietal aroma of Xinomavro

It is well documented that varietal aroma is an important parameter of wine quality. Chemical compounds responsible for wine varietal aroma are sourced from secondary grape metabolites. Until today little research is conducted on the influence of vine clone on the grape aromatic content of Greek grape varieties. Xinomavro (Vitis vinifera L.) is one of the most important Greek grape varieties, valuable for the wine industry of Northern Greece since it contributes to the production of PDO wine of Naoussa, Amindeo and Goumenissa.

Physical-chemical and sensory characterization of wine made with the cultivar syrah produced in a double pruning system

In recent years, the consumption of fine wines in Brazil has increased significantly, a phenomenon that is also reflected in the expansion of production to new regions. In the brazilian southeast for example, the so-called “winter wines” are being produced, through management in two cycles, one of formation and one of production, with two prunings and one harvest per year, a technique known as double pruning, with vineyards established at altitudes close to or above 1,000 m above sea level.

Closure permeability modulates the aroma expression of monovarietal white wines during bottle ageing

Bottle ageing is a critical period for wine quality, as it undergoes various chemical and sensory changes during storage. Ideally, a phase of qualitative ageing, during which wine sensory quality improves, is followed by a decline of quality. Understanding how different oenological variables influence these phases is a key challenge in modern winemaking. Recent studies highlighted the significant role of oxygen in modulating reactions involving volatile and non-volatile components, impacting aroma evolution during bottle aging. Oxygen exposure of wine during bottle ageing is mediated by closure.

From a local to an international scale: sensory benchmarking of PDO wines. Quincy and Reuilly PDO wines (Sauvignon blanc) as a case study (France)

In a collective marketing strategy, the Protected Designation of Origin (PDO) can be used as a quality indicator. To highlight terroir specificities, it is useful to know how the wines are positioned on the local, national or international market from a sensory point of view. This is especially true for a comparison of varietal wines (e.g. Sauvignon blanc). We focus on the case of two closed Loire Valley PDO (France): Quincy and Reuilly. Three distinct tastings were organized. Firstly, at the local level comparing the 2 PDO (11 and 9 wines, 17 professional assessors); secondly at a regional level adding 3 closed PDO: Menetou-Salon, Sancerre and Pouilly-Fumé (3 wines per PDO, 16 assessors) and thirdly at an international level comparing these 5 PDO with Sauvignon Blanc wines coming from South Africa, New Zealand and Chile (1 to 3 wines per PDO, 19 assessors). All the wines were from the 2019 vintage and were considered to have a traditional elaboration process without contact with oak. A sensory descriptive analysis was performed using an aroma wheel allowing to combine a Check-All-That-Apply methodology, often used in sensory benchmarking, with a hierarchical structuration of the attributes. The aim is to facilitate data acquisition in a professional context without common training, to consider the hierarchical relationships among the attributes during the data analysis and to be able to characterize wines with a large range of sensorial variability. We use univariate, multivariate and clustering analyses. Similarities and differences between Quincy and Reuilly PDO wines and other Sauvignon blanc wines were identified. Specific attributes can distinguish the two PDO and different proximities exist with other local PDO, while clear differences were observed compared to international wines. Our study contributes to propose and discuss a method to do a wine sensory benchmarking highlighting sensory specificities linked to origin.