Drone-Enabled Data Collection for Sustainable Viticulture: A Practical Framework for Precision Viticulture

Drone-Enabled Data Collection for Sustainable Viticulture: A Practical Framework for Precision Viticulture Sofia Samoili1[0000-0002-1997-007X], João Filipe Gaspar2, Renato Ferreira3, Maria da Graça Vieira Duque Carreira Monteiro1, Pedro Sebastião2[0000-0001-7729-4033], Sofia Kalakou1[0000-0003-0992-2303] 1 Instituto Universitário de Lisboa (ISCTE-IUL) - Business Research Unit (BRU), Avenida das Forças Armadas, 1649-026, Lisbon, Portugal 2 Instituto Universitário de Lisboa (ISCTE-IUL) - Department of Digital Technologies (DTDA SINTRA), Avenida Heliodoro Salgado 3, 2710-569 Sintra, Portugal 3 Instituto Universitário de Lisboa (ISCTE-IUL) - Department of Social and Enterprise Sciences (DCSE SINTRA), Avenida Heliodoro Salgado 3, 2710-569 Sintra, Portugal Abstract. A structured framework is developed for the use of unmanned aerial vehicles (UAVs), particularly drones, in precision viticulture, within the SmartVitiNet project. The approach includes a detailed process for drone operation in vineyards, incorporating technical, environmental, and operational parameters, from equipment setup and pre-flight route planning for efficient data collection, to post-flight data management. Route automation (area and waypoint-based), and advanced sensor and imaging configurations, allow the collection of precision data for vineyard monitoring, and the rapid coverage of large areas. The drone-based framework enhances sustainability through datadriven insights, which optimise pesticide, fertiliser, and water use. Moreover, the guide presents a replicable methodology that facilitates a wider adoption of drones by vineyard managers, bridging the gap between AI-assisted practices and agricultural reality. The uses and benefits of drones in viticulture are also thoroughly discussed. By integrating geospatial data collection with climatesensitive flight planning, the methodology contributes to the broader digital transition towards sustainability land and resource management. The study provides a representative potential of the future of drone systems services as mobile data platforms for climate-wise agriculture, aligning with incorporating innovation for improving sustainability in transport, mobility, and environmental performance in agriculture. Keywords: Precision viticulture, Unmanned Aerial Vehicles (UAVs), Drone system services, Mobile data platforms, Sustainable agriculture. 1 Introduction Precision viticulture (PV) is a precision farming approach applied to vineyard management that employs advanced technologies. The aim is to monitor and optimise 2 S. Samoili, J. Filipe Gaspar, R. Ferreira, M. Monteiro, P. Sebastião and S. Kalakou production efficiency, quality and profitability, while minimising environmental impact. Sensors, GPS, satellite imagery and Unmanned Aerial Vehicles (UAVs) are frequently used tools. UAVs, else commonly known as drones, enable precise monitoring, detailed data collection and analysis, leading to more informed decisionmaking and optimised resource allocation in vineyard operations. The adoption of drones in viticulture may address significant challenges that the agricultural sector faces currently. These include the impacts of climate change, such as high temperatures and irregular rainfall, as well as the rising production costs associated with product waste, and sustainability concerns related to the environmental impact of traditional machinery, the overuse of fertilisers and pesticides. Drones appear to be a promising solution to mitigate these issues by enabling a more targeted and hence efficient vineyard management. Multiple studies outline the use of UAVs in precision viticulture or more generally in precision agriculture. In Beloev [2016] the advantages and disadvantages of UAVs in precision agriculture among other sectors are discussed, and advanced information technology solutions are explored to enhance their effectiveness. An overview of the potential of UAVs in enhancing intelligent transportation systems (ITS) and smart city infrastructure is presented in another recent study [Lucic et al., 2023]. UAV challenges and solutions during deployment are mentioned, including energy limitations, mission scheduling, and fleet coordination. The importance of efficient scheduling frameworks and charging station planning to ensure successful UAV operations in urban environments is also highlighted. Spachos and Gregori [2019] discuss the integration of wireless sensor networks and UAV platforms for real-time monitoring of the vine growth environment. This enables on-demand imaging and high-resolution data collection, optimising production and input application in a cost-effective manner [Spachos & Gregori, 2019]. Gavrilović et al. [2024] explore the integration of UAVs and artificial intelligence (AI) in precision viticulture, focusing on vine detection and vineyard zoning. With UAV imagery and YOLO deep learning algorithm for vine detection, 90% accuracy is achieved. For vineyard zoning, K-means algorithm is applied to geospatial data, including NDVI and nutrient content assessments. This enables efficient resource management tailored to each zone's needs [Gavrilović et al., 2024]. UAVs are also used for analysing vineyard variability through data collection for row segmentation and disease detection [Sassu et al., 2021]. Broader agricultural applications show improvement by the use of UAVs. Smart crop management using digital agriculture tools such as IoT, remote sensing, and AI are discussed in Fuentes-Peñailillo et al. [2024]. These technologies when applied to viticulture, improve production efficiency and sustainability, through monitoring, field surveys and crop growth tracking. Singh et al. [2024] presents applications of drone technology in providing solutions that increase agriculture productivity and crop quality, including crop health monitoring, weed management, evapotranspiration estimation, spraying, seed sowing, and growth assessment. The integration of IoT and deep learning using UAVs for farm management is explored to enhance decisionmaking processes in agriculture, through UAVs with sensors and AI algorithms to monitor farming operations, improving efficiency and productivity [Mishra, 2023]. Practical Framework for Drone-Enabled Precision Viticulture 3 The potential of UAVs to facilitate precision agriculture in improving environmental performance, reducing costs and supporting vineyard management, is depicted in the literature review. However, practical UAV integration into actionable viticultural operations remains fragmented. To address this gap, a structural operational framework for UAV use in vineyards is developed through the SmartNitiNet (Smart and Sustainable Drone-Assisted Viticulture Excellence Network) project. The guide provides a practical, replicable process for drone operations, from equipment setup and pre-flight planning to route automation, post-flight data management and analysis. Standardised checklists and procedures are freely available on the Competence Centre website of the project 1 , making advanced technology more accessible and feasible for agricultural practices, and fostering sustainable viticulture. Section 2 details drones and sensor technologies in viticulture, Section 3 explains the guide to drone operations in vineyards, and the applications and benefits of implementing this drone-based framework follows in Section 4, to conclude with Section 5 with the practical limitations and the future potential of drone technology in viticulture. 2 Drones and Sensor Technologies in Viticulture The core technologies for an optimised and sustainable vineyard management include the use of UAVs as remote sensing platforms. They operate by using optical sensors to collect qualitative and quantitative data about crops or other objects from distance. UAVs outweigh satellite or manned aerial imaging, as they allow frequent and ondemand data collection, as well as highly localised monitoring under specific environmental or phenological conditions. The data collection is achieved by measuring the emitted, transmitted or reflected electromagnetic radiation by the objects on the ground, namely in this case the grapevines. By detecting specific wavelength ranges, or else bands, the UAV platforms can be used for the detailed analysis of vegetation, crops, water status, pests and diseases. Their capacity to operate with many bands and at low altitudes ensures a high resolution spatial imagery for detailed vineyard-level assessment, which may detect the early onset of issues, such as disease, nutrient deficiencies, or water stress. 2.1 Types of Sensors A variety of sensors can be mounted on UAVs, enhancing their versatility, as they can be changed depending on the required monitoring tasks for the vineyard management. The main sensor types that can be selected according to the application are summarised in Table 1. Table 1. Main sensor types for viticulture management with UAVs Sensors Details Uses RGB Capture light in visible spec- - Field mapping and surveying 1 https://smartviticc.iscte-iul.pt 4 S. Samoili, J. Filipe Gaspar, R. Ferreira, M. Monteiro, P. Sebastião and S. Kalakou trum (R, G, B) - Visual inspection - Growth & yield monitoring estimation Multispectral Detect up to 15 wavebands incl. NIR - Crop health monitoring - Soil and irrigation analysis - Normalized Difference Vegetation Index (NDVI) computation - Disease detection Hyperspectral Over 100 continuous bands - Crop health monitoring - Soil analysis - Yield prediction Thermal Infrared Detect emitted heat (IR) - Irrigation monitoring - Plant health monitoring - Soil condition analysis - Frost protection LiDAR Light Detection and Ranging Pulsed laser light - Canopy and soil analysis - Terrain mapping - Yield estimation - Disease detection ER & EMI Electrical Resistivity (ER): Measure electrical conductivity or resistivity of soil. Electromagnetic Induction (EMI): Use electromagnetic fields to detect conductivity variations of the ground. - Soil conditions (moisture, salinity) - Groundwater mapping - Soil composition and structure for soil health analysis RGB sensors are used in agricultural monitoring to capture visible light spectrum data (red, green, blue) used for field mapping, visual inspection, and estimating crop growth and yield. Multi-spectral sensors capture data from specific spectral bands, including those outside the visible range, assisting in monitoring crop health, analyse soil and irrigation status, compute vegetation index, such as NDVI, and detect diseases. Hyper-spectral sensors have finer detail data for assessing crop health, soil conditions, and predicting yields. Thermal Infrared sensors detect emitted heat and measure temperature variations across the vineyard, data that are used for identifying waterstressed areas, comprehensive monitoring of plant health, irrigation needs, soil conditions, and frost protection. LiDAR (Light Detection and Ranging) sensors use pulsed laser to measure distances and create 3D maps of the terrain and vegetation. They are used for precise terrain mapping and yield estimation, for canopy and soil analyses, and disease detection. Electrical Resistivity (ER) & Electromagnetic Induction (EMI) sensors provide insights into soil moisture, salinity, and environmental conditions, vital for managing irrigation and monitoring soil quality. Each type of sensor has specific applications that assist in improving precision agriculture and sustainable vineyard practices. The integration of these technologies with a geospatial information system (GIS) and data analytics provides a more comprehensive understanding of a vineyard’s conditions, hence allowing targeted inter- Practical Framework for Drone-Enabled Precision Viticulture 5 ventions and optimised yields. Nevertheless, operational, regulatory, and data-related issues prevent UAVs adoption from viticulturists. To address these issues and introduce an effective use of drone data, an operational framework has been developed within the SmartVitiNet project to guide practitioners in using of drones for the benefit of a more sustainable viticulture. 3 Methodological Framework for Drone-Based Vineyard Monitoring 3.1 Operational Framework Basis and Overview A structured guide has been developed in the framework of SmartVitiNet, a project funded by the European Union’s European Innovation Council and SMEs Executive Agency (EISMEA), to introduce and facilitate the use of drones in viticulture. It was designed to ensure reproducibility, as well as regulatory and safety protocols, following an iterative development approach that incorporates field trials’ feedback from viticulturists, drone operations and technology providers. The suggested standardised procedures aim to improve viticulture operations through the use of the advanced drone technologies. The guide includes visual technical procedures explanations and recommended actions to facilitate safe, efficient, and replicable drone operations for vineyards applications. It is divided in three modules: the (i) setup process, (ii) the data collection process, from equipment preparation for the drone flight to post-flight data management, and (iii) the overall uses and benefits of drones in viticulture. It supports both areabased and waypoint-based route automation, and incorporates sensor configuration guidance for different monitoring objectives, such as canopy assessment or disease detection. Further guidance and training materials are available through the SmartVitiNet Competence Centre website and a global video distribution platform 2 , which offers demonstrations and tutorials for safe and efficient UAV operations, and increases the project’s outreach. 3.2 Setup Process To ensure a safe and successful drone flight a thorough preparation is required prior to the day of the flight. This pre-flight setup process involves certain steps to be conducted for a drone configuration suitable for data collection in viticulture. First standardised checklists are used to guide the setup and ensure consistency across operations. Flight routes are created subsequently to ensure a comprehensive data collection, addressing the operations’ objectives and terrain characteristics. All equipment is prepared in advance, namely drone hardware, sensors, and batteries. The timing of the flight should be also scheduled according to the solar zenith to ensure optimal lighting conditions and replicability across the data collection. Weather forecasts are consulted to maximise visibility, and minimise wind and precipitation conditions. Finally, a 2 https://www.youtube.com/channel/UCuTJC9pcMeqJ8YzX8rkTQeA 6 S. Samoili, J. Filipe Gaspar, R. Ferreira, M. Monteiro, P. Sebastião and S. Kalakou verification of equipment readiness is performed prior to the flight. Following this structured process ensures that the equipment is properly prepared, and that a comprehensive set of necessary safety and operational checks are completed. Therefore, data collection is optimised and relevant risks are minimised. The steps are further elaborated as follows. Checklists. Comprehensive checklists are essential for a safe and successful operation. These are divided in four main phases: a) Before the Flight Day: The operation goal is defined during this phase. No-fly zones are identified, the flight route is planned, and the required permissions are secured. Following this, the weather conditions are reviewed, and the drone undergoes a full inspection, namely firmware and apps are updated, and all batteries are charged to ensure optimum performance. b) Pre-start on Flight Day: Flight readiness and operational safety of the site and the airspace, are ensured in this phase with on-site checks, and by identifying potentially suitable emergency landing sites. All personnel involved in the operation are briefed on the operation plan, and their roles. Then the drone is inspected, by checking the airframe, battery levels. Finally, the IMU and compass are calibrated to ensure accurate flight control and navigation. c) Start-flight: It includes the just before the flight and during take-off actions. A stable GPS is confirmed by lock, and the signal connection between the drone and controller is checked for strength. The designated flight route should is selected and verified, and the take-off area is cleared of any obstacles. As the motors engage, the drone is monitored for any unusual sounds that could indicate mechanical issues. System notifications are continuously monitored for battery status, signal interference, wind conditions, or any alerts that may affect flight safety. d) Finish Flight: Post-flight procedures are described to monitor the flight-to-home sequence and ensure the designated landing area is clear of any physical obstacles. Once the drone is landed, the drone and controller are turned off safely. A final inspection of the equipment is then conducted, to check for any signs of damage or wear. Finally, the collected data are reviewed to ensure operation objectives were met and that the files have been properly saved for post-processing. Route Planning. Comprehensive data collection relies on efficient route planning. Two distinct flight routes must be created, namely an area-based route for top-down ortho-imagery collection (90° angle), and a waypoint-based route for angled (oblique) imagery with customisable parameters at each point, which is allows the replication of the operation. Based on the vineyard characteristics, the operator selects the type of route through a flight planning software. Spatial and spectral resolution precision is achieved through parameterisation of flight characteristics (altitude and speed), alongside the adjustment of image overlap and ground sampling distance (GSD). Equipment Preparation. Following route definition, the day before the flight, all drone equipment must be thoroughly prepared for the operation. This involves con- Practical Framework for Drone-Enabled Precision Viticulture 7 firming that the necessary equipment is available and fully charged for the duration of the operation, ensuring that the drone's propellers are in proper working order, and verifying that the memory card is empty or has the required storage for the upcoming data collection. It includes also the firmware updating on both the remote control and the drone to the latest versions for optimal performance and safety. Flight Timing. Aiming to obtain high quality data, the flight should be scheduled under optimal illumination conditions, namely during a time that the shadows are minimised and the light is maximised uniformly across the vineyard. The typical chosen hour is one hour before and one hour after the sun’s highest point (solar zenith), which varies according the location. A consistent zenith time for all operations should be maintained to enhance comparability between datasets. This improves also the accuracy of reflectance-based vegetation indices, such as NDVI (Normalized Difference Vegetation Index). Weather Conditions Assessment. To avoid data inconsistencies caused by environmental variability, guarantee safety and avoid equipment damage from rain or high speed wind, the weather forecast must be confirmed before the flight. A reliable source’s hourly forecast can be used for this purpose. Operators should verify that no adverse conditions are expected during the operation. Final Equipment Confirmation. A final verification step concludes the pre-flight setup, by verifying operational conditions and data integrity prior to the flight, alongside an equipment check prior to take-off. This step is crucial to confirm that drone and remote control batteries are fully charged, memory card is correctly installed in the drone, and that all equipment is working before storing them in a carrying case for transport to the flight location. 3.3 Data Collection Process To ensure the reproducibility and high accuracy of a data collection operation in viticulture, a second operational phase is outlined in the framework for drone operations, which addresses the on-site data collection process. It is conducted in four steps, encompassing the flight and the post-flight stages. The equipment setup is the initial step prior to take-off, aiming to ensure proper functionality of the drone and avoid potential damage. The drone and controller are prepared for the flight, verifying their operation. The Real-Time Kinematic (RTK) module should be then attached to the drone to enhance its geolocation accuracy, while the protective covers are removed to prevent obstruction to the gimbal, and the motor arms are unfolded to avoid wear on the motors. For an accurate flight and navigation, a clean and flat surface should be found, and proceed with geolocation and calibration of internal sensors routine, synchronising with GPS satellites to ensure positional accuracy. 8 S. Samoili, J. Filipe Gaspar, R. Ferreira, M. Monteiro, P. Sebastião and S. Kalakou To streamline the pre-flight preparation, during calibration and geolocation procedures, the route setup can be initiated. The drone’s trajectory is selected among the pre-created flight routes, and critical for safety parameters are adjusted. This includes the Return-to-Home (RTH) altitude to navigate safely over obstacles, the obstacle avoidance distances to prevent collisions with surrounding objects, the maximum altitude to limit the drone within legal and safe requirements, and the battery warning levels customisation to prevent unexpected power loss. Prior to the final flight mission upload to the drone, the mission summary is reviewed and all parameters are confirmed, namely estimated flight duration, distance, number of waypoints and number of camera captures. During the flight, the drone’s behaviour and stability is monitored continuously, for operational safety and data consistency issues. Observing the drone's visual behaviour, including its stability and response to controls, helps the early detection of potential issues, such GPS signal loss, or mechanical failures. Real-time data, such as altitude, battery level, GPS signal strength, and wind speed, provides essential insights into the drone's operational status, allowing the operator to make informed decisions. To ensure data precision, camera feedback allows for quality control of imagery. A safe conclusion to the flight is secured by monitoring the automated RTH and landing phases, until the drone descends and lands safely, allowing for manual intervention to prevent collision risks. The post-flight procedure begins with the deactivation of the drone’s power systems and inspection for any damage to its structure or propellers, to avoid future performance and safety issues. The packing of the equipment follows, in reverse order to the setup procedure, to ensure an efficient setup for future flights. 4 Applications and Benefits of Framework in Sustainable Viticulture The SmartVitiNet methodological framework for drone-based vineyard monitoring with high-resolution, geo-referenced imagery, offers viticulturists, vineyard managers and technology providers a structured, replicable and safe method for collecting and processing data for the advancement of sustainable viticulture. Precision technology benefits vineyard management decisions, as viticulturists can inspect large areas with improved accuracy and time efficiency. This ensures an optimised use of resources, such as water, fertilisers and pesticides, by applying them based on the specific needs of different areas of the vineyard. As drones with optical sensors can serve as remote sensing platforms, collecting imagery of crops with very high spatial resolution, their integration in vineyard management provides impactful benefits in production costs, efficiency and sustainability. Among the most impactful applications is variable rate application. Spatial variations may appear in vineyards’ soil and plant composition, indicating variable needs across different areas. Drones equipped with multispectral or thermal sensors can identify specific needs, allowing growers to optimise use of resources, as pesticides, fertilisers and water can be administered to specific areas in a field according to the require- Practical Framework for Drone-Enabled Precision Viticulture 9 ments. This approach presents a dual advantage, as the released volume of chemicals and water waste are reduced, as well as the traditional machinery use, decreasing thereby the emitted pollutants. From an economic and operational perspective, the integration of drones in viticulture following a standardised framework, could reduce costs as the effectiveness of the process is expected to improve yield predictability, reduce crop losses and disease outbreaks with timely detection of water stress and nutrient deficiencies. The operational framework can be readily integrated into vineyard management, as its simplicity allows its adoption even from viticulturists with limited technical expertise. In addition, the framework demonstrates adaptability and replicability across various vineyard scales and geographic locations, as it has been demonstrated in pilots taken place in four European countries. Knowledge transfer across the viticulture sector is effectuated also through the SmartVitiNet Competence Centre, where wine producers, technology providers, authorities, and researchers can find training materials, relevant policy and technological information. This contributes to the European agenda for digital and sustainable transformation in agriculture. 5 Conclusions, Challenges and Future Outlook This study presented a structured and replicable framework for the integration of drones in precision viticulture, which was developed in the framework of the EUfunded SmartVitiNet project. The framework standardises the operational procedures of UAV monitoring in viticulture, aiming to facilitate their adoption by vineyard managers and enabling them with a technically feasible approach, from setup and data collection to post-flight. Furthermore, the proposed framework contributes to bridging operational efficiency and sustainability, by promoting through the drone-assisted monitoring of viticulture the targeted use of water, fertilisers and pesticides. Beyond the promotion of an environmental responsible vineyard, the method also creates the link between a sensor-based analytical tool and decision-support systems. Apart from the demonstrated benefits, the integration of UAVs in viticulture presents certain technical and regulatory challenges. Technical constraints of the operation of UAVs in viticulture are the primary limitations, among which are the weather and terrain conditions, and equipment endurance. Wind, precipitation and sunlight exposure affect image quality, while battery life dictates the flight time. Training and certification of the drone operator may discourage small producers from using drones, as aviation regulations require licensed pilots and details operational planning in European areas. The high-resolution imagery storage and handling constitutes a practical challenge, which requires relevant digital infrastructure and significant storage capacity. Therefore, scalable support systems and simplified tools should be presented to small and medium enterprises to overcome this limitation. Future work will focus on integrating UAV data with additional sources, such as IoT ground sensors, and machine learning analysis, in order to create more compre-