Wildfires are a critical challenge exacerbated by climate change and combustible biomass accumulation. The objective of FireTwin is to provide predictive wildfire-risk maps to enable forest managers and public authorities to implement preventive actions that mitigate the risk of extreme events. FireTwin addresses the lack of reliable fuel load and moisture information by fusing satellite imagery (Sentinel-2) with climate data to estimate Live Fuel Moisture Content (LFMC). The product delivers high-resolution risk maps and short-term forecasts through interoperable API data services or a ready-to-use hosted platform, supporting operational decision-making and climate resilience.
Customers and their Needs
The main customer segments encompass public administration (wildfire prevention and civil protection services), timberland owners, forest managers, insurers/reinsurers, and carbon project developers.
These customers face the challenge of obtaining localised, accurate, and continuous fire-risk intelligence for decision-making. Their needs include better preparedness during critical risk windows, auditable operational alerts and reducing the high technical cost of integrating new tools. FireTwin addresses these points by providing risk intelligence and short-term forecasts (24 to 72 hours) to plan patrols, assess eligibility and pre-position emergency resources. Furthermore, its dual delivery model (API or web) reduces technological friction during user onboarding.
Targeted customer/users countries
Spain (beachhead market), the rest of the Iberian Peninsula and Mediterranean Europe countries.
Product description
FireTwin is a modular cloud-native system (PaaS) that integrates Earth observation (EO) data, meteorological networks (AEMET, ECMWF) and in-situ data with artificial intelligence (AI) modules to generate Live Fuel Moisture Content (LFMC) modelling.
In terms of its architecture, the system features automated ingestion modules, an AI engine that calibrates wildfire risk (nowcasts and 24 to 72-hour forecasts) and European cloud infrastructure, guaranteeing data sovereignty.
To facilitate customer interaction, the service relies on a dual delivery architecture: an API gateway and interoperable OGC geospatial services for direct integration into control centres or third-party systems and an intuitive web platform providing interactive maps, dashboards and alert configurations for non-technical users.
Added Value
Compared to competitors and existing tools (such as Copernicus EFFIS/GEFF or Technosylva), the added value of FireTwin lies in its hyper-local focus on current conditions (operational nowcast). It offers high spatial detail and low data-to-map latency (≤6 hours processing time), which is critical for operational responses.
Unlike competitors that require complex IT integrations or cover overly generic areas, FireTwin stands out for its lightweight integration via interoperable APIs and fully explainable risk outputs and alerts. All of this is complemented by localised support and auditable historicity, an essential requirement for governance reporting and insurance or carbon market audits.
Current Status
The activity advances through two interconnected cycles: a de-risking cyclefocused on validating and confirming the feasibility of the short-term forecast capability (meteorological integration and LFMC modeling), and a product development cycleaimed at building and scaling the comprehensive operational service to a TRL 7 level. The system builds on mature technological heritage from over a decade of prior research and development by Agresta in satellite forest monitoring (including projects like FIREPOCTEP+, VIS4FIRE, and CILIFO), moving from fragmented tools to a centralised, cloud-native architecture.
Agricultural stakeholders need accurate and timely information, which current EO-based monitoring services cannot fully provide, as they mainly focus on past and present conditions.
The more predictive insight is available on crop stress dynamics, the more effective decisions can be made to reduce risks and optimise operations across the agri-food value chain. Therefore, the objective is to develop an EO- and AI-based predictive decision-support platform that provides short-term (two to six weeks) forecasts of multi-stressor crop risks.
Agricultural input suppliers and insurance companies primarily require high-resolution, parcel-level information for operational decision-making. Their needs are addressed by integrating multi-source data, including Sentinel EO imagery, meteorological data and farm-level datasets, and delivering products such as predictive stress maps, risk indicators and early warning alerts. At the same time, stakeholders such as insurers, integrators and financial institutions require scalable information at regional to national level. For these users, the platform provides aggregated risk indicators, continuous monitoring and forecast-based analytics to support planning, reporting and risk management.
Customers and their Needs
The team assessed customer needs through direct engagement and analysis of operational workflows across the agri-food value chain. The results show that stakeholders require not only predictive information, but decision-ready, use-case-specific insights that can be embedded into daily business processes.
Primary users — agricultural input suppliers and insurance companies — expressed several concrete, recurring needs. Input suppliers require early-season crop emergence detection at parcel level to replace delayed and often inaccurate statistical data, enabling earlier demand forecasting and more efficient logistics planning. They also need field-specific advisory support, as current recommendations are based on generic assumptions, regional weather data, or farmer feedback, which reduces credibility and limits adoption.
In addition, suppliers seek targeted identification of stress-affected sub-parcels, allowing them to optimise product placement (e.g. fertilizers or crop protection) while remaining compliant with increasingly strict EU sustainability regulations. Another key use case is objective verification of crop emergence for emergence-based insurance schemes, which is currently manual, slow and costly.
Insurance companies, on the other hand, require objective, scalable indicators for early loss anticipation, particularly two to six weeks before damage materialises. Today, their processes rely largely on historical data and reactive field inspections, which limits their ability to price risk accurately or manage reserves efficiently. They also need spatially explicit stress indicators to support parametric insurance products and to reduce uncertainty in claims assessment.
Secondary users — agricultural integrators and financial institutions — have complementary needs at larger scales. Integrators require regional yield and quality forecasts to reduce procurement uncertainty and avoid contract shortfalls, while banks need aggregated, predictive indicators of climate and production risks to improve agricultural credit risk assessment and reduce exposure to non-performing loans. Across all segments, a common requirement emerged: a solution that integrates multiple data sources (EO, weather, soil and farm data), provides short-term forecasts and translates them into actionable, workflow-integrated insights rather than standalone analytics.
Targeted customer/users countries
The service primarily targets agricultural input manufacturers and distributors, as well as insurance companies and brokers in Hungary.
Secondary target users include agricultural integrators and banks and financial institutions.
Geographically, the service is initially focused on Hungary, but it can be extended to Central and Eastern European countries with similar agro-climatic conditions and crop structures. In the longer term, the solution is scalable to broader European markets where Earth Observation data and agricultural datasets are available.
Product description
EnviMAP Agro is an online geographic information system (GIS)-supported platform based on Sentinel satellite imagery, meteorological data, soil moisture information and proprietary farm-level datasets to support agricultural risk management and decision-making activities.
The system integrates multi-source EO and in-situ data with AI-based models to detect and predict crop stress impacts. The online platform offers several products, such as multi-stressor risk maps (drought, heat stress, pests and diseases), short-term (two to six weeks) predictive forecasts, crop development indicators, and an alerting service providing early warning alerts on emerging and forecasted stress events at parcel to regional scale, as well as scenario-based analyses.
In addition, the system provides aggregated risk indicators and monitoring services to support operational planning and risk assessment across the agri-food value chain. The online platform enables effective dissemination of information and user-friendly interfacing with customers, without special knowledge or infrastructure need (fat server/thin client model). Clients can check and manage their data for areas of their interest, 24/7 from anywhere there is internet access. Especially for corporate users and decision makers, dashboards and reporting tools are available to access key information immediately.
Added Value
The innovation resulting from the project enables predictive, short-term (two to six weeks) assessment of multi-stressor crop risks, providing accurate and actionable information on drought, heat stress, pests and diseases. Accordingly, large agricultural areas can be analysed in a short time with low human resource requirements, supporting simultaneous evaluation of crop conditions at parcel to national scale.
By using the service, input suppliers can reduce input misuse by 10–15%, lowering costs and supporting compliance with EU sustainability regulations. At the same time, early-season forecasts improve logistics and distribution efficiency, reducing overstock and enabling more targeted product placement.
For insurance companies, predictive indicators support accurate pricing and the development of parametric insurance products. The system enables early risk mitigation, reducing payouts by up to €150–300/ha in stress years and improving loss ratios. EO-based validated risk maps also enable faster and more cost-efficient claims handling and more efficient reserve capital allocation.
Remote sensing data combined with AI models allows continuous monitoring and forecasting of crop conditions. The system provides precise spatial information on stress occurrence, supporting targeted interventions and improved decision-making across the value chain. With the implementation of continuous monitoring and forecasting, up-to-date information is available on evolving risks. If stress levels exceed defined thresholds, alerts are generated to support timely response by users and decision makers. This leads to improved production stability, strengthened supply chain resilience and reduced financial exposure, including potential reduction of non-performing loan ratios in agricultural portfolios.
Current Status
The kick-off of the project took place on 07 April 2026. Following the kick-off, activities have been initiated to collect and consolidate user and system requirements through structured stakeholder interactions, focusing primarily on input suppliers and insurance companies. The results of this phase will be used to refine key functionalities, define system architecture and prioritise use cases.
The backbone of the platform builds on Envirosense Hungary’s existing developments, including the GeoRISK system used by agricultural insurers and previously developed EO-based crop monitoring and classification methodologies. These provide a validated starting point for predictive modelling and platform development.
Initial activities include the assessment of available EO, meteorological, soil and farm-level datasets, as well as the definition of preliminary system architecture, data flows and module interfaces. In parallel, key abiotic and biotic stressors are being identified and prioritised based on user relevance and modelling feasibility. The next milestone will focus on the integration of data sources and the development of predictive modelling workflows, including multi-stressor detection and short-term forecasting capabilities, as well as their implementation within the online platform.
Farming, land management and flood planning all depend on understanding the environment, yet keeping track of change across large areas is difficult. The data these sectors need sit in separate systems that do not interact with each other. Ground sensors are precise, but they cover only a small area. Satellites cover huge areas, but they cannot see conditions on the ground in detail.
Anganode closes that gap. It combines satellite imagery with data from ground-based sensors to give a single, reliable picture of what is happening in a given place and uses each source to check the other.
The activity focuses on three areas, combining ground sensors and satellite data to make the difference: using water more efficiently, measuring carbon more confidently and forecasting flood risk.
The platform is built for people who are not remote-sensing specialists. Through a simple point-and-click interface, an application programming interface (API) or a developer kit, teams set up their own monitoring without writing code or hiring Earth observation experts. The result is faster answers at a lower cost. Anganode speeds up environmental insights by up to 70% and cuts reporting costs by 50% to 80% compared with traditional consultancy work. It also supports carbon and environmental reporting, so organisations act earlier on risks such as drought, flooding, wildfire and land instability.
Customers and their Needs
Anganode serves organisations across Europe and North America that need to monitor the environment at scale. Most face the same barrier: traditional satellite analysis needs Earth observation specialists to interpret raw imagery, which leaves teams with fragmented data, slow delivery and blind spots. Anganode focuses on three areas, where combining ground sensors and satellite data makes the biggest difference.
Irrigation efficiency. By tracking how much water is leaving a field and where it is overwatered, Anganode helps growers decide when to irrigate, top up or hold off. Combining evapotranspiration readings from ground devices and satellites replaces guesswork based on generic weather patterns with clear, timely guidance on when to water.
Carbon measurement. Land-management choices, such as planting or protecting vegetation, affect whether carbon is stored or released. Anganode helps measure that impact more confidently and supports Monitoring, Reporting and Verification (MRV) workflows. Verifying stored carbon accurately, for example, and proving how much carbon a stand of new trees has stored, is genuinely difficult. The activity is working towards evidence robust enough to support that process.
Flood risk. Anganode improves the forecasting, prediction and live assessment of flood extent by combining local sensor data with satellite observations. Better local inputs sharpen satellite-based estimates, giving communities and operators earlier, clearer information to act on.
The same fusion approach extends to further use cases. For example, it supports monitoring wildfire risk, ground movement and vegetation growth around critical power infrastructure, helping network operators act before an outage occurs.
Targeted customer/users countries
Environmental monitoring organisations in Ireland, Germany, the Netherlands, France, Belgium and the United States.
Product description
Anganode runs on a cloud-native platform built on AWS and Kubernetes. It treats ground sensor data and satellite imagery as two independent sources, then blends them for each use case.
Collect. Data from ground sensors arrives over standard protocols (MQTT, CoAP, AMQP) or low Earth orbit satellite networks and moves through an Apache Kafka data pipeline. Separately, users request the Earth observation layers they need from satellite constellations such as Sentinel and Landsat, which the platform downloads automatically and shows alongside the sensor data.
Choose. Users select the sensors and satellite outputs relevant to their use case, without needing to be an Earth observation expert or data scientist.
Blend. The two sources are combined into machine-learning models built for each use case, such as evapotranspiration and watering schedules for irrigation or flood forecasting.
Assist. Built-in generative AI, using AWS Bedrock, on-device language models and a secure knowledge base, helps users interpret results, make decisions and work through the initial setup.
Output. Results appear in a standard web browser as interactive 2D and 3D maps of the terrain, overlaid with environmental data and configurable time-series charts.
An orchestration layer underpins all of this. It is designed for flexibility, hiding the specialist image processing so non-experts can still make a decision backed by data. In principle, it can support any use case. The activity begins with irrigation efficiency, carbon measurement and flood risk.
Added Value
Most alternatives fall into one of two fields. Some hand over raw satellite imagery that still needs in-house experts to interpret. Others lock you into fixed, single-purpose hardware. Rivals in data fusion tend to build one-off, consultancy-led setups for a single use case, which are hard to adopt and harder to scale. Anganode takes a different approach.
Vendor-neutral software. It works with any sensor, satellite mission or manufacturer, built on a proven industrial IoT platform.
Self-calibrating and cross-checked. Broad satellite observations are checked against precise local ground readings, and the two approaches validate each other. This quality control across the ground and satellite views makes results both more accurate and more trustworthy.
Faster to set up. Preparing a new monitoring workflow drops from around 11.5 days to roughly 3.5 days.
Reporting support. It produces structured records to support carbon and environmental reporting, including workflows aligned with Monitoring, Reporting and Verification (MRV) and disclosure rules such as the EU CSRD.
Current Status
The activity began on 16 July 2026 and is in its early stages.
Twinspector is designed to provide reliable and independent access to very high-resolution satellite imagery that helps utilities better understand and manage critical infrastructure and the environment.
The product addresses a growing need for timely, accurate and scalable information to monitor assets such as railways, power networks, pipelines and surrounding land areas, especially as infrastructure systems become more sensitive and exposed to environmental and operational risks.
The primary objective of Twinspector is to deliver consistent and very high-quality Earth observation data that enables early detection of changes, supports risk assessment, and improves decision-making for infrastructure operators and public authorities.
By offering both standard and three-dimensional stereo imagery, the product allows users to gain deeper insights into terrain, vegetation and asset conditions over large geographic areas.
Customers and their Needs
Twinspector addresses the needs of infrastructure operators and public-sector organisations that rely on continuous, accurate and scalable geospatial information to manage spatially distributed assets.
Primary customers include operators of linear infrastructure such as railways, electricity transmission and distribution networks, pipelines and transport corridors, as well as national and EU authorities responsible for infrastructure oversight, environmental monitoring, and regulatory compliance.
Infrastructure operators require regular access to very high-resolution optical imagery to detect small-scale changes, assess risks, and monitor asset conditions along narrow corridors over large geographic areas. Stereo imagery and three-dimensional information are essential to support height-related risk assessment and vegetation-related hazards.
Customers also require predictable revisit rates, consistent imaging conditions, and reliable long-term data availability to integrate satellite-based insights into operational planning and maintenance workflows.
Across all customer groups, seamless integration of data products into existing geospatial systems and long-term service continuity are key requirements for sustainable operational adoption.
Targeted customer/users countries
Europe (including Germany and wider EU/EEA), United Kingdom, North America (USA/Canada), and selected growth markets (e.g., Asia-Pacific, including Australia and Africa).
Product description
Twinspector is a dedicated Earth observation mission based on two identical optical satellites flying in a trailing formation, enabling the acquisition of very high-resolution mono and stereo imagery for infrastructure and environmental monitoring.
The product offers panchromatic imagery with sub-meter spatial resolution (≤ 50 cm GSD) and multispectral imagery with meter-scale resolution (≤ 2 m GSD) across four spectral bands (blue, green, red, near-infrared). Stereo imaging is enabled through controlled off-nadir acquisitions with defined stereo geometry, supporting the generation of three-dimensional information products, such as Digital Surface Models, with high geometric consistency. Typical swath widths of > 14 km allow efficient coverage of narrow corridors and long linear assets.
Intelligent onboard processing clips imagery to the corridor Area of Interest and applies cloud masking to reduce downlinked data and accelerate delivery.
On ground, the processing chain produces analysis-ready products (including stereo-derived 3D information) that feed LiveEO’s risk analytics services (Treeline and SurfaceScout). Users interact via a web application and API, with export formats (e.g., geospatial layers and reports) for integration into enterprise asset and work management systems.
Added Value
Twinspector strengthens the value of LiveEO’s existing analytics products, Treeline and SurfaceScout, by providing a dedicated and reliable source of very high-resolution mono and stereo satellite imagery tailored to infrastructure monitoring needs. Treeline supports utilities and railway operators in identifying vegetation-related risks along linear assets, while SurfaceScout focuses on detecting third-party activities, terrain instabilities, and other threats in the surrounding environment.
By ensuring predictable data availability and consistent imaging conditions, Twinspector significantly increases the scalability of these products, allowing customers to monitor larger networks more frequently and with greater confidence.
By reducing dependency on manual field inspections and emissions-intensive aerial patrols, Twinspector helps lower operational costs, improve safety for field personnel, and support more sustainable infrastructure operations. As a dedicated and scalable data source, Twinspector strengthens the long-term reliability and continuity of LiveEO’s services, enabling customers to embed satellite-based insights into routine asset management workflows.
Current Status
The Twinspector activity is in the definition phase. User needs, mission objectives, and system-level requirements are consolidated and aligned with LiveEO’s existing products and target use cases. Initial concepts for the space and ground seg-ments, including stereo imaging and data processing, are defined, providing a solid basis for progressing into detailed design and implementation.
Current work focuses on system engineering, risk management, and procurement planning for critical payload and platform elements, including long-lead items. Coor-dination between the prime (service and mission owner), platform supplier, payload supplier and electronics partner is established to ensure consistent design and veri-fication approaches across the space segment.
The request for new satellite-based observations and data management is largely dependent on centralised ground-based operations and constrained by the availability of the ground station, where data and commands are exchanged. These data are only made available for further analysis after many hours.
This limits the possibility to implement approaches such as ‘tip and cue’, since the ground segment needs to plan monitoring acquisitions, wait until the data reach a ground station, and only then process and analyse the data. In case something interesting is present in the acquisition (e.g. wild fire, dark vessel or illegal phishing), a follow-up acquisition can then be planned to gather more data, but only hours later.
The large delay between these actions limits the effectiveness of the monitoring actions. High-performance computing units and inter-satellite link connections through satellites are planned for many future constellations, and exploiting those assets will dramatically change the operational scenario: Earth observation data can be immediately processed, with AI algorithms used to extract relevant features and support autonomous decisions through AI agents, just using on-board resources and allowing an almost real-time monitoring, which would be impossible otherwise solution using transmitters based on vacuum tubes, like TWT, or a set of discrete solid-state components.
Customers and their Needs
Leonardo acts as the anchor partner and reference early adopter, with direct involvement in defining the system needs and mission integration constraints, particularly in view of its future Earth observation constellation.
In parallel, Leonardo has initiated preliminary technical discussions with Satlantis (Spain) and D-Orbit (Italy):
Satlantis, a vertically integrated smallsat operator and payload manufacturer, has expressed their interest in testing on-board autonomy features for responsive Earth observation scenarios. They represent a relevant user profile due to their dual role as integrator and operator of high-resolution optical payloads.
D-Orbit, as a platform and in-orbit service provider, is evaluating TASCNET for integration on its ION Satellite Carrier as part of future mission architectures with embedded edge computing.
Targeted customer/users countries
Spain and Italy
Product description
The TASCNET product will integrate the typical capabilities of the control ground segment with autonomous capabilities allowed by the growing availability of high-performance computing resources installed on the satellite platform and the possibilities offered by AI algorithms and agents.
The main goal of the product is to migrate some of the control functionalities currently performed the ground segment to the space segment. This change allows a constellation with required capabilities to autonomously update the mission plan if specific conditions are found, enabling the possibility to implement complex scenarios, such as ‘tip and cue’, without the need of support from ground operators, greatly reducing the latency.
TASCNET combines the elements from two main components into a hybrid mission control:
On-board software components. These components are installed on satellites with computing resources able to run complex algorithms required to process the payload data, identify the main features within the image, autonomously decide the following operations, and operate with other satellites of the constellation.
On-ground control ground segment components. These components have the responsibility to build the mission plan for each satellite of the constellation, but also to control the complex algorithms and the criteria for more complex tasks, such as ‘tip and cue’..
Added Value
TASCNET activities focus on three main pillars:
On-board scheduling and autonomous tasking: it enables satellites to autonomously generate observation schedules and dynamically adjust them in orbit to react to new information, ensuring optimal use of platform resources without ground intervention.
On-board Earth observation processing: TASCNET includes a library of AI models able to process payload data directly in orbit, extracting relevant features (e.g. changes, targets, anomalies). This unlocks new use cases such as intelligent alerting and automated re-tasking.
Next-generation ground segment: traditional control segments must evolve to integrate hybrid planning logics. TASCNET supports this evolution with a multi-master planning architecture that brings together autonomous onboard planning with strategic ground-based oversight.
We foresee the following main added values:
On-board EO Processing and Autonomous Planning
Detects opportunities and anomalies in real time, adjusting plans without waiting for ground commands.
Enables ultra-low-latency reaction chains, such as ‘tip and cue’.
Reduces bandwidth consumption and operational overheads by prioritising only high-value data for downlink.
AI-based Mission Optimisation
Intelligent agents.
Observation planning to improve responsiveness and reduce redundancy.
Hybrid Ground Segment Adaptation
TASCNET enables coexistence of on-board and on-ground control logic.
Enhances user interaction with the mission via predictive dashboards and dynamic plan reconciliation tools.
Current Status
The project has started identifying the most relevant constellation characteristics that can enable added value services, such as ‘tip and cue’ without ground segment assistance, the most relevant use cases and the associated mission planning challenges. The requirements review has been successfully held, and the team is working on the preliminary design review preparation.
X-band High-Efficiency High-Gain 18-W Power Amplifier, for synthetic aperture radar (SAR) antennas;
X-band High-Efficiency 30-W High-Power Amplifier (HPA), for Earth observation satellite data down-link.
These products have the aim to offer a highly efficient, green and economical solution to system integrators of TR modules for SAR antennas. Being the HPA the most power-consuming of the RF line-up, the target is to develop a monolithic HPA, with the embedding of the driver function, leading to a huge increase of the power efficiency and a saving of up to 40% of the power consumption of SAR antennas.
This strongly increases the competitiveness of Earth observation satellites companies, which will offer a meaningful extension of the service to end users. In addition to this HPA, with the same objective of saving energy and room, a fully monolithic and very efficient HPA in X band for the data down-link is under development, in replacement of the current solution using transmitters based on vacuum tubes, like TWT, or a set of discrete solid-state components.
Customers and their Needs
The main targeted customers are Earth observation satellites, specifically the SAR antennas, and airborne and ground-based AESA radar integrators.
The product provides a solution to their need of better performances, higher integration and cost reduction with respect to the actual exploited solutions.
The main end users and potential customers provided their support for the definition of the product main requirements.
Targeted customer/users countries
The main users/customers are from Europe (Italy, France, Germany).
Product description
The product consists of a chipset of advanced High-Power Amplifier MMICs capable to provide the optimisation of the next X-band SAR Antennas for Earth observation in terms of power consumptions and encumbrances. In detail, two different monolithic circuits are proposed:
An X-band High-Efficiency High-Gain 18-W Power Amplifier, to be exploited as the final element of the RF frontend in the TR modules for SAR antennas.
An X-band High-Efficiency 30-W High-Power Amplifier, to replace the SSPA currently used on Earth observation satellites for data downlink.
When optimised for the maximum efficiency operating point, the GaN technology selected for the proposed activity is credited to deliver an output power of about 3.5 W/mm.
Based on these characteristics, a preliminary analysis identified the following baseline architectures for the two MMICs, respectively, targeting the required levels of Output Power, PAE and Gain.
The following figures show the preliminary architectures defined for the two HPAs.
18W pulsed X-band HPA: Baseline architecture
30W CW X-band HPA: Baseline architecture.
Added Value
The breakthrough for both amplifiers, with respect to the current solutions, is the exploitation, at X-band, of the 0.15 µm GaN/SiC technology usually optimised for applications at 30 GHz.
The high available gain, as well as the improved efficiency and the usual high-power density level of Gallium Nitride processes, allow to reach the required output power with a very high efficiency, leading to a notable reduction of both power consumption and dissipation.
Furthermore, the well-known robustness of this technology gives the advantage to satisfy the tight reliability constraints required by space applications with limited scaling of the overall performances.
For the two specific cases, the request of the High-efficiency Amplifier comes from the need to replace 1-to-1 the MMIC currently used in the RF frontend of the X-band SAR TR modules, but with the objectives to significantly reduce the power consumption, operating with a very-high efficiency, and, at the same time, to remove the buffer amplifier now used to amplify the power at the output of the beamformer, possible by integrating in the proposed single MMIC up to 4 stages of gain.
For a direct comparison, with respect to the amplifiers used on the TR modules of COSMO-SkyMed Second Generation (CSG), the proposed solution allows twice the power (18 W) with no more than 30% of additional power consumption (33 W).
FieldFinder addresses the challenge of generating consistent agricultural field boundary maps across diverse geographic regions. Existing field delineation models are often trained on regionally limited datasets and show reduced performance when applied to agricultural landscapes with different field structures.
FieldFinder extends an existing field boundary detection model with additional ground-truth data from multiple global regions. This improves model generalisation and reduces geographic bias in delineation performance.
The activity integrates multi-regional training datasets and scalable processing workflows using satellite imagery, primarily from Sentinel-2. The resulting service provides globally consistent agricultural field boundary datasets. These datasets support applications including supply-chain transparency, sustainability certification, agricultural monitoring, and policy support. Customers access the product through data delivery, API integration or tailored analytics. The activity builds on previous ESA-supported developments and extends them into a globally applicable field boundary mapping solution.
Customers and their Needs
Target customers include sustainability certification bodies, supply-chain transparency platforms, agricultural monitoring providers, public institutions, and climate and environmental analytics companies. These organisations require consistent agricultural field boundaries to monitor production areas, assess land-use change, and support regulatory compliance.
Existing field boundary datasets often perform well in specific regions but show reduced performance when applied to agricultural landscapes with different field structures, crop types and management practices. This limits the scalability of monitoring solutions and increases manual mapping efforts.
Customers require globally consistent field boundary datasets that perform reliably across diverse agricultural environments. FieldFinder addresses these needs by improving model generalisation through multi-regional training data and scalable processing workflows.
Customers engage through pilot projects, validation exercises, and early adoption of the generated datasets.
Targeted customer/users countries
Worldwide, with primary focus on the EU, the UK, and Switzerland as key customer and decision-making regions. Additional target markets include major agricultural production areas in South America (especially Brazil and Colombia), Southeast Asia (Indonesia and Malaysia), South Asia (India and Vietnam), and selected African countries such as Kenya, Ghana, Uganda, and Egypt, where demand for scalable agricultural monitoring and land-use transparency is growing.
Product description
FieldFinder is a satellite-based AI service that extracts agricultural field boundaries at scale. The system processes Earth observation imagery, applies deep learning segmentation models, and generates structured field boundary datasets.
The product architecture includes:
Satellite data ingestion (Sentinel-2 and complementary datasets)
Data preprocessing and tiling
Deep learning field boundary detection
Post-processing and geometry optimisation
Validation using ground-truth datasets
Data delivery via API and geospatial formats
The activity focuses on improving model generalisation by integrating multi-regional training datasets. This reduces geographic bias and improves performance across diverse agricultural landscapes.
Customers interact with the product through geospatial data delivery, API access, or tailored analytics depending on integration requirements.
Added Value
FieldFinder improves the consistency of agricultural field boundary mapping across different geographic regions. Existing solutions often rely on regionally limited training datasets and therefore show reduced performance when applied to new agricultural environments.
By integrating multi-regional training data, FieldFinder reduces geographic bias and enables globally consistent field delineation. This supports scalable applications in sustainability monitoring, supply-chain transparency, land-use analysis, and agricultural monitoring.
The activity builds on previous ESA-supported developments and extends them into a globally applicable service. This reduces development risk and increases technical maturity.
The resulting datasets reduce manual mapping effort, improve consistency, and enable large-scale monitoring across diverse agricultural landscapes.
Illustration: Field delineation comparison in the same area. Left: model trained on different region data. Right: FieldFinder with improved boundary detection. Credit: Marple
Current Status
FieldFinder builds on an existing field boundary detection model and datasets developed in previous activities. Initial models and ground-truth datasets are available and provide a solid foundation for extension to additional geographic regions and agricultural landscapes. The activity expands multi-regional training datasets, improves model generalisation, and validates performance across selected pilot regions. Processing pipelines and system architecture are operational at prototype level and undergo iterative refinement. Customer engagement and pilot use cases support validation, performance assessment, and alignment with user requirements.
Satellites generate petabytes of raw data every day, yet traditional downlink-then-process workflows create delays and drive mission cost. TychoBoB solves this bottleneck with a general-purpose, FPGA-based edge computer that is installed directly inside a small satellite.
The platform is built for ease-of-use and affordability, enabling research institutes, SME payload developers and mission owners to run custom, high-performance applications in orbit with only a few engineering hours.
This is achievable with:
Remote access & hardware leasing.
Seamless external IP integration.
Developer-friendly toolchain.
Customers and their Needs
Early-stage Earth observation mission owners, university smallsat labs, and SME payload/algorithm developers need an affordable way to validate high-throughput onboard processing before they lock in a satellite bus or raise full mission funding.
Today, they either buy costly engineering models late in the schedule or rely on software emulators that hide hardware bottlenecks. Both options create re-design risk, months of delay and significant extra cost. TychoBoB addresses these gaps by letting a two-or-three-person team load its own Linux services onto a flight-representative board in days, instead of months. A cloud-hosted FlatSat and short-term EM leasing remove the up-front CAPEX barriers, while a built-in functionality and Yocto flow reduce integration surprises later. During this InCubed activity, NTNU (research) and S&T Norway (SME) act as pilot users, feeding real requirements, IP cores and test cases into the development process.
Targeted customer/users countries
Norway, Sweden, Denmark, Finland, Germany, Netherlands, France, Spain, United Kingdom, Italy, Greece, Portugal, Belgium, Austria, Ireland, and Switzerland.
Product description
TychoBoB is a Xilinx UltraScale+-based FPGA/ARM edge computer tailored for small satellites.
The compact board hooks directly to a wide range of sensors via LVDS/CameraLink, 1 Gb Ethernet, CAN, UART and SPI, ingests high-rate data streams, and executes sophisticated AI/ML workloads to detect events and alert end-users with minimal latency.
Key innovations:
Developer-friendly toolchain – reproducible Yocto build flow, automated CI tests and step-by-step tutorials that cut setup from weeks to hours..
Remote access and hardware leasing – prototype on a cloud-hosted FlatSat or with a leased engineering model before committing to flight hardware.
Seamless external-IP integration – architecture refined through continuous user input from algorithm developers, ensuring painless drop-in of third-party processing IP.
Added Value
TychoBoB cuts mission cost at two critical points: (1) it slashes the amount of raw data that must be downlinked, and (2) it lets operators process data in-orbit instead of paying for large cloud clusters on the ground.
A 2025 third-party market study comparing a “traditional” Earth observation satellite stack with a TychoBoB-enabled stack found:
Data-download charges: A decrease from € 230 000 to € 115 000 per satellite, an estimated 50 % reduction.
Ground IT processing & hosting: A decrease from € 50 000 to € 5000, an estimated 90 % reduction.
Overall system-hosting cost: From € 300 000 to € 250 000.
Total estimated savings are around € 210 000 per satellite, per year, without sacrificing data quality or latency.
Beyond direct OPEX gains, TychoBoB delivers:
Faster revenue-to-orbit: remote access and leasing mean payload teams iterate faster.
Lower technical risk: reproducible Yocto builds, automated CI and proven IP-integration pipelines cut integration failures.
Greater mission flexibility: users can push new AI models after launch, keeping the satellite relevant over its lifetime.
These combined advantages translate into shorter payback periods and higher ROI for commercial, governmental and research missions.
Current Status
As of July 2026, the activity is progressing towards a pilot test in October. Hardware is being sent to production and the software development is nearing an MVP. In terms of customer interest, the product is garnering increasing amounts of attention. Several large customers are interested, despite the original business hypothesis. However, they require additional radiation tolerance updates to hardware and software. The original smaller customer segments are also interested. TychoBoB is a likely candidate for two upcoming IOD opportunities. The deadline for go/No-go for these opportunities await confirmation.
A digital platform that combines GEOSAT’s Very High-Resolution Earth Observation satellites with other data sources to provide up-to-date visualisations and insights for local government city management. It offers two main product levels:
City digital models: High-resolution (40 cm) imagery from GEOSAT-2 is used to create detailed and accurate urban representations for GIS-based operations. Depending on subscription, users can access varying coverage areas and update frequencies, potentially reaching intra-daily updates. A 3D version is also available by integrating satellite imagery with in-situ data and stereo imaging techniques.
Tailored management insights: Delivered as a value-added service on the same platform, this level uses AI/ML and automated processing to provide advanced applications. These include land use classification and change detection, route optimisation for large vehicles and emergency scenarios, and monitoring the health of green urban areas.
Customers and their Needs
The platform primarily serves municipal authorities, supporting urban planning, cadastre, mobility, civil protection, and green area management. It has been already used in collaboration with multiple municipalities in Portugal and Spain.
It also targets intermunicipal and regional authorities, including regional coordination and development bodies, autonomous governments, and cartographic institutes, as well as administrations in island and southern regions. National mapping agencies are key users, such as geographic institutes, land and territory management directorates, and national statistics and geography organisations. Engagement is also beginning with national public administration entities.
The platform enables continuous land use monitoring, including tracking urban expansion, detecting changes in residential, industrial, and service areas, and mapping green spaces for distribution, preservation, and carbon capture potential. It also provides detailed road infrastructure characterisation, including route classification and critical point identification, while supporting cadastre management and assessing urban intervention impacts.
Key challenges addressed include maintaining up-to-date urban classification, optimising routes for heavy and non-standard vehicles, automating green space health assessment and monitoring sustainability through carbon capture evaluation, supporting more efficient and sustainable urban management.
Targeted customer/users countries
The approach is to first demonstrate the value proposition and project outcomes through a pilot project with Portuguese municipalities, followed by engagement with the private sector. In a subsequent phase, the target audience will expand to the Iberian Peninsula, including Portuguese regional and national authorities, as well as Spanish regional and national authorities. In the next step, the geographical scope will extend to Europe and Latin America, targeting regional and national governments. Finally, international expansion is foreseen, enabling engagement with regional authorities worldwide.
Product description
GEOSAT is developing a dynamic urban operations management tool delivered via a digital platform that integrates its Very-High-Resolution (VHR) Earth Observation satellites with other data sources, providing up-to-date visualisations and actionable insights for city management. Its first phase focuses on short-term priority needs of municipal services.
The platform has two main components:
City digital models: Using GEOSAT-2’s 40 cm resolution imagery, it generates detailed, accurate urban representations for GIS operations. Subscription options offer variable coverage and update frequency, with potential intra-daily updates via the full Atlantic constellation. 3D models are available by combining satellite imagery with in-situ data through stereo-pair techniques.
Tailored management insights: Delivered as a value-added service, these AI/ML-powered applications cover land use classification, change detection, pathway optimisation for large vehicles and non-standard scenarios, and green space health assessment.
Building on existing VHR imagery and processing methods, the platform will evolve to integrate new satellite data, automate visualisation, use satellite and in-situ data, and adapt models to municipal-specific needs, including customised land use categorisation and quantification of green areas. It addresses critical problems: inconsistent or outdated information, lack of easily accessible tailored insights, inability to model scenarios, and scattered data from multiple sources.
Added Value
The GEOSAT platform is a single, integrated solution combining multiple data sources with advanced tools to analyse information and generate actionable insights. It addresses customer-specific needs through a new product focused on visualisation and insight generation, with short-term capabilities like land use classification, change detection, pathway optimisation, and green space assessment, and medium- to long-term developments such as 3D model generation, scenario simulation, and pollution monitoring.
The platform directly tackles the challenges of urban data analytics, the requirements for connected urban data, and the operational needs of local governance. Unlike traditional GIS tools or static mapping services, it provides recurrent monitoring of urban environments, enabling real-time detection of changes in land use, infrastructure, and green spaces.
Its flexibility and scalability through subscription-based coverage and update frequencies, including potential intra-daily updates, ensure municipalities have the most current and precise data. The integration of 3D city models and fusion of satellite with in-situ data allows simulation of interventions, optimised routing for large vehicles, and assessment of non-standard scenarios.
By centralising fragmented data and delivering tailored, automated insights, the platform reduces operational workload, accelerates decision-making, and supports sustainable urban management, offering a level of precision, timeliness, and operational relevance unmatched by existing solutions.
Portuguese municipalities manage complex territorial challenges, including urban sprawl, wildfire risk, invasive species, water quality and regulatory compliance, but lack the technical capacity to access and interpret the Earth observation data that could inform their decisions. Most of Portugal’s 308 municipalities have no GIS specialists as staff.
LandOS addresses this gap with an AI-powered conversational platform. Municipal staff ask questions in natural Portuguese, such as “How has urban sprawl changed in our municipality over the last five years?”, and receive evidence-based answers in the form of maps, charts, and downloadable PDF reports. The platform combines Sentinel-2 satellite imagery with Portuguese national datasets (CAOP administrative boundaries, COS land cover, REN/RAN ecological reserves) through an agentic AI system that autonomously selects and executes the appropriate geospatial analysis.
The project develops, deploys, and validates this platform in a 12-month product development cycle, piloting with three Portuguese municipalities (Fundão, Odemira, Mértola) and targeting TRL 7 readiness for commercial launch.
Customers and their Needs
Target users: Municipal staff in Portuguese municipalities, specifically:
Urban planning technicians (‘técnicos de urbanismo’): need to monitor urban expansion, validate development permits against REN/RAN protected zones, and support PDM (municipal master plan) revisions with evidence
Environmental officers (‘técnicos de ambiente’): need to track water quality in reservoirs, monitor invasive species spread, and report on environmental indicators
Civil protection coordinators: need seasonal fire risk assessments cross-referenced with ICNF wildfire records to prioritise prevention actions
Municipal managers (‘gestores’): need high-level dashboards and reports for decision-making and compliance reporting
Key challenges: These users are domain experts but not GIS specialists. They cannot use traditional remote sensing tools. They need information delivered in Portuguese, in formats they already understand (reports, maps, simple dashboards), without requiring training in satellite data interpretation.
Targeted customer/users countries
Country: Portugal (pilot municipalities: Fundão, Odemira, Mértola). Expansion target: Southern European municipalities
Product description
LandOS provides a cloud-based Software-as-a-Service (SaaS) platform accessible through a web browser. The system has three main layers:
Conversational Interface – Municipal staff type questions in natural Portuguese. The system understands municipal terminology and context (e.g., parish names, regulatory references, land-use categories).
Agentic AI Engine – An AI based orchestration system automatically determines which data sources and analytical tools are needed, retrieves the relevant data, executes the analysis, and synthesises the results into a clear Portuguese-language response with supporting maps and charts.
Analytics & Visualisation – Five core analytics modules are delivered as part of the minimum viable product: urban sprawl detection (COS change analysis), water quality monitoring (Sentinel-2 spectral indices), fire risk assessment (multi-factor: NDVI, topography, weather), invasive species detection (Random Forest ML on Sentinel-2), and regulatory compliance checking (REN/RAN spatial intersection). Results are displayed as interactive maps, charts, and branded PDF reports.
The platform is built on Project Zeno, a proven open-source geospatial AI framework, and runs on cloud infrastructure that ensures EU data residency and GDPR compliance.
Added Value
LandOS relies on Copernicus Sentinel-2 multispectral imagery as its primary Earth Observation data source, complemented by the European Digital Elevation Model (EU-DEM) from the Copernicus Land Monitoring Service.
Sentinel-2 enables four of the five core analytics: water quality monitoring through NDWI and chlorophyll-a spectral indices; fire risk assessment through NDVI vegetation health analysis; invasive species detection through spectral signature classification; and urban sprawl analysis through multi-temporal land cover comparison. Without satellite-derived data, these analyses would require costly and infrequent field surveys that most Portuguese municipalities cannot afford.
The space added value is the ability to provide consistent, repeatable, municipality-scale environmental monitoring at a frequency (every 5 days with Sentinel-2) and spatial coverage (all 308 Portuguese municipalities simultaneously) that no ground-based alternative can match. By combining this EO capability with AI-powered natural language access, LandOS removes the technical barrier that has historically prevented municipalities from benefiting from the Copernicus programme.
Current Status
The project kicked off in March 2026 following contract signature with ESA under the InCubed “EO for Municipalities” call. The Kick-Off Meeting took place on 24 March 2026.
The first development phase (Months 0–2) focused on building the foundations of the platform: the cloud environment that hosts the service, the secure sign-in system for municipal users, and the full Portuguese localisation of the user interface and AI agent.
Three pilot municipalities (Fundão, Odemira, Mértola) are confirmed for the activity through signed Letters of Support. The first on-site visit took place in Odemira on 7 May 2026, and included a programme of interviews with municipal staff across urban planning, environment, civil protection and management. On-site visits to Fundão and Mértola are scheduled in the coming months. LandOS also attended the Portugal Smart Cities Summit in Lisbon (12–14 May 2026).
The Requirements Review, the first contractual milestone, took place on 9 June 2026.
