Terroir 2020 banner
IVES 9 IVES Conference Series 9 International Terroir Conferences 9 Terroir 2020 9 History and innovation of terroir 9 Detection of spider mite using artificial intelligence in digital viticulture

Detection of spider mite using artificial intelligence in digital viticulture

Abstract

Aim: Pests have a high impact on yield and grape quality in viticulture. An objective and rapid detection of pests under field conditions is needed. New sensing technologies and artificial intelligence could be used for pests detection in digital viticulture. The aim of this work was to apply computer vision and deep learning techniques for automatic detection of spider mite symptoms in grapevine under field conditions. 

Methods and Results: RGB images of grapevine canopy attacked by the spider mite (Eotetranychus carpini Oud) were manually taken in commercial vineyard (Etxano, Basque Country, Spain) under natural day light conditions. Leaf segmentation in images was performed based on computer vision techniques, isolating target leaves with spider mite visual symptoms from the vineyard canopy. HSV colour space was used to consider colour variations representing symptoms on the leaves, separating these values from those of saturation and brightness of the image. Spider mite detection was done using Convolutional Neural Networks (CNN) models with an artificially augmented dataset for the classification of leaves with this pest symptoms. An accuracy surpassing 75% was obtained using a hold-out validation.

Conclusions: 

High accuracy proves the effectiveness of the trained model in the classification of grapevine leaves. Computer vision techniques were useful to image classification on the relevant pixels. Additionally, deep learning techniques provided a robust model to find complex features of spider mite visual symptoms.

Significance and Impact of the Study: Non-invasive technology and artificial intelligence shown promising results in the automatic detection of pests in commercial vineyards.

DOI:

Publication date: March 23, 2021

Issue: Terroir 2020

Type: Video

Authors

Inés Hernández1, Salvador Gutiérrez2, Sara Ceballos1, Ignacio Barrio1, Fernando Palacios1, Ana M. Diez-Navajas3, Javier Tardáguila1*

1Televitis Research Group. University of La Rioja, 26007 Logroño, Spain 
2Department of Computer Science and Engineering, University of Cádiz, 11519 Puerto Real, Spain 
3Department of Plant Production and Protection, NEIKER-Basque Institute for Agricultural Research and Development, Basque Research and Technology Alliance (BRTA), 01192 Arkaute, Spain

Contact the author

Keywords

Deep learning, computer vision, pests, grapevine, crop protection

Tags

IVES Conference Series | Terroir 2020

Citation

Related articles…

The Douro region: wine and tourism

The Demarcated Douro Region (DDR) dates from 1756, when it was recognized as one of the first demarcated regions in the world. The DDR economic activities fit the terroir model and are based on wine and tourism.

Grape ripening delaying with combined use of leaf removal and natural shading in Manto negro (Vitis vinifera L.) under deficit irrigation

The increasing frequency of heat waves during grape ripening presents challenges for the production of high-quality wine grapes. This underscores the significance of developing effective irrigation and canopy management techniques to optimize both yield and grape quality.
A field experiment was carried out during 2021 and 2022 using Manto negro wine grapes to study the effect of two irrigation strategies and different light exposure levels on grape quality. In a four-block experimental vineyard at Bodega Ribas in Mallorca, two irrigation treatments—moderate and severe deficit irrigation—were implemented. Within each irrigation plot, three light exposure treatments were randomly assigned, encompassing exposed clusters from pea size, non-exposed clusters, and shaded clusters after softening.

Some applications come from a method to concentrate proteins

All techniques usually used to assay proteins was not reliable in vegetable extract due to interferences with the components included in extracts like polyphenols, tanins, pectines, aromatics compounds. Absorbance at 280nm, Kjeldhal assay, Biuret and Lowry methods, Acid Bicinchonique technique and Bradford assay give the results depending on the composition of extract, on the presence or not of detergent and on the raw material (Marchal, 1995). Another difficulty in these extracts for the quantification of proteins comes from the large amount of water included in vegetable and the low concentration of proteins. Thus in red wines, proteins are usually not taken into account due to their low concentration (typically below 10 mgL-1) and to the presence of anthocyanis and polyphenols.

Sensory profiles of Shiraz wine from six Barossa sub-regions: a comparison between industry scale and standardised small lot research wine making

Aims: The Barossa wine region in South Australia comprises six sub-regions and is renowned for its Shiraz wines. However, there is no comprehensive documentation of the distinctive sensory characteristics of wines from these sub-regions.

Quantification of the production of hydrogen peroxide H2O2 during wine oxidation

Chemical studies aiming at assessing how a wine reacts towards oxidation usually focus on the characterization of wine constituents, such as polyphenols, or oxidation products. As an alternative, the key oxidation intermediate hydrogen peroxide H2O2 has never been quantified, although it plays a pivotal role in wine oxidation. H2O2 is obtained from molecular oxygen as the result of a first cascade of oxidation reactions involving metal ions and polyphenols. The produced H2O2 then reacts in a second cascade of oxidation to produce reactive hydroxyl radicals that can attack almost any chemical substrate in wine.