Fraunhofer Center for Assistive Information and Communication Solutions – AICOS
30.06.2023
Technology has become an essential tool in modern agriculture, especially with the use of Artificial Intelligence (AI) in vineyards and other agricultural applications.
Producers and entrepreneurs in the Demarcated Douro Region (DDR) now have at their disposal a mobile technology with AI for monitoring pests in vineyards, which is one of the biggest threats to the Region's wine sector. With this innovative solution, pest management can now be optimised through a system that automatically detects key insects and provides recommendations to combat the threats identified. A preliminary version of the solution is being made available by GeoDouro, free of charge, to winemakers. Market entry is planned for 2024.
Telmo Nogueira, Manager of the Design and Development Department of GeoDouro, acknowledges “producers are very enthusiastic about this solution. The possibility of providing a system that helps them monitor and prevent pests, as well as a more direct contact with specialists who can improve pest monitoring, is seen with great interest by farmers”. The impact of this technology on the routines and daily lives of producers, but above all, on the quantity and quality of grapes and wine, is undeniable. As the expert from GeoDouro puts it: “through this solution, it was possible to computerize, automate and standardize the entire process of monitoring and preventing pests that, increasingly and in a context of climate change, affect the cultivation of vines.”
But the vineyards are not the limit! As this solution “is built in a scalable way, it will be possible to apply the solution to new pests not only in vineyards, but also in other crops, such as olive groves, fruit and vegetables, where in the latter we already have an initial monitoring for tomato moth”.
The pest monitoring solution was developed and validated with five stakeholders. Currently, and according to Telmo Nogueira, GeoDouro “is expanding the solution to new stakeholders, not only in the Douro Region but also in other wine regions, in order to optimize the solution and improve some components.” Once the improvement phase is complete, the technology will be marketed and commercialised, which is planned for 2024.
The Technology: from scratch to market
The identification and counting of insects is conventionally done by winemakers through visual inspection, a time-consuming and tiring manual process, limited by their lack of taxonomic knowledge. The great climatic heterogeneity in the DDR, together with the consequences of climate change, aggravate the situation. Environmental factors such as temperature influence the development cycle of insects, and so far there is no systematic collection of meteorological data from each trap. Over three years, the technology has been developed, tested and sharpened. Around 9,000 insect images were collected and annotated, and AI-based computer vision techniques were applied, allowing the automation of the pest detection and counting process, making it more efficient and traceable.
The system integrates three components:
(1) a mobile application for the winemakers, for use in the field, in order to automatically recognise and count pests in traps, and record the phenological state;
(2) a central web module, to centralise and store the information collected by the viticulturist, and
(3) a web portal for taxonomy experts, to observe the results of the automatic monitoring of insects in the different plots, being able to remotely support the viticulturists in their pest management.
The solution, developed by Fraunhofer Portugal AICOS (FhP-AICOS), began to be designed in 2020, under the name of EyesOnTraps*, in partnership with the Association for the Development of Duriense Viticulture (ADVID) and GeoDouro - Consultoria e Topografia, Lda. Participating as stakeholders representing the wine sector were Sogevinus Quintas SA, Adriano Ramos Pinto - Vinhos SA and Sogrape Vinhos, SA.
“The EyesOnTraps solution is an excellent example of how AI is reshaping agriculture processes through digitalization and automation”, explains Luís Rosado, Senior Researcher at Fraunhofer Portugal AICOS (FhP-AICOS) and Project Manager of the developed solution. “The technology has the potential to improve current pest management procedures in viticulture, from time spent on monitoring tasks and support key-insects identification and counting, to promote a more optimized and sustainable usage of phytosanitary treatments.”
EyesOnTraps is designed to help vineyard owners and managers identify the presence of plagues. The system uses a network of smart traps equipped with sensors that collect data on the plagues' behavior, such as flight patterns and population density. This data is then analyzed using AI algorithms, which can identify patterns and anomalies that may indicate the presence of pests.
One of the primary benefits of the EyesOnTraps system is its ability to detect pests early, before they can cause significant damage to the crop. Early detection allows vineyard managers to take preventive measures, such as spraying pesticides or deploying other pest control methods, before the pests have a chance to reproduce and spread. This not only helps to protect the crop but also reduces the need for more intensive and costly pest control measures later on. Another benefit of the EyesOnTraps system is its efficiency. Traditional methods of pest monitoring often involve manual inspections of the vineyard, which can be time-consuming and labor-intensive. The EyesOnTraps system, on the other hand, can monitor multiple locations simultaneously and provide real-time data on pest activity. This can save time and resources, allowing vineyard managers to focus on other important tasks.
“The solution emerged from substantial multidisciplinary research from our Human-Centred Design, Intelligent Systems and Connected Things groups, which showcases how our three FhP-AICOS’ groups work as one”, comments Luís Rosado.
Surely, a solution to keep an eye on!
This article was published in our LinkedIn Newsletter. Subscribe it to stay up to date with our work.
Co-funded by:
*EyesOnTraps, a SI R&DT project (NORTE-01-0247-FEDER-039912), implemented from 2020 to 2023, represented an investment of over 400 thousand euros.