Periodic Reporting for period 1 - EARTHPULSE (Satellite Processing By Artificial Intelligence - Enabling Satellite Analytics)
Reporting period: 2023-07-01 to 2024-06-30
Climate change poses a significant threat to global infrastructure, food security, and economies, requiring proactive approaches to mitigate impacts. Traditional emergency protocols are typically reactive, resulting in delayed damage assessments and increased socio-economic impacts. There is a pressing need for proactive vulnerability assessments to identify the most vulnerable areas and assets, enabling adaptive responses and minimizing damage before events occur.
Despite substantial investments in the satellite data market, such as the €8.2 billion allocated to the Copernicus Satellites, the full potential of satellite data remains untapped due to the complexity and high cost of data handling. This limits market penetration and return on investment (ROI). EarthPulse aims to accelerate the uptake of Earth Observation (EO) data by integrating Artificial Intelligence (AI), reducing costs, and making satellite analytics accessible to all users. Our solution, SPAI (Satellite Processing by Artificial Intelligence), is specifically designed to overcome existing barriers in the EO sector.
SPAI Innovations
• User-Friendly Interface for Non-Experts: Utilizes Natural Language Processing (NLP) to enable non-experts to easily extract actionable information from EO data without requiring specialized knowledge.
• Advanced Interface for Experts: Offers a Domain Specific Language that allows data scientists to interact with EO data and AI models in near real-time, enhancing data manipulation capabilities.
• Cloud Optimized GeoTIFFs (COGs): Facilitates efficient, affordable cloud workflows, enabling seamless data storage, access, and processing on cloud platforms.
SPAI Procedural Innovations
1. Simplified EO Data Access and Processing: SPAI simplifies the handling of EO data, removing traditional barriers to access.
2. AI-Ready Data Preparation: Employs the SCAN labeling tool to create datasets specifically designed for training neural networks.
3. Open PytorchEO Library: Supports the design, training, and export of neural networks, making it easier for developers to work with AI models.
4. Model Repository: Provides a platform for the operational use of uploaded models, fostering collaboration and sharing within the community.
SPAI revolutionizes the EO processing chain, making it iterative and continuously learning, thereby reducing training costs and improving scalability.
Expected Impact
1. Proactive Disaster Management: SPAI enables adaptive responses to climate extremes, significantly reducing socio-economic damages by allowing timely interventions.
2. Enhanced Market Penetration and ROI: By increasing the uptake of EO data, SPAI enhances the ROI for investments in the space economy, encouraging further investment.
3. Increased Accessibility and Usability: SPAI democratizes satellite analytics, making it accessible to a broader range of users, including those without technical expertise.
4. Improved Scalability and Efficiency: SPAI enhances the scalability and cost-effectiveness of EO data processing, making it suitable for a wide range of applications and industries.
Overall, EarthPulse aims to transform the EO landscape by making satellite data analysis more accessible, efficient, and impactful, addressing critical climate change challenges and market opportunities.
Despite substantial investments in the satellite data market, such as the €8.2 billion allocated to the Copernicus Satellites, the full potential of satellite data remains untapped due to the complexity and high cost of data handling. This limits market penetration and return on investment (ROI). EarthPulse aims to accelerate the uptake of Earth Observation (EO) data by integrating Artificial Intelligence (AI), reducing costs, and making satellite analytics accessible to all users. Our solution, SPAI (Satellite Processing by Artificial Intelligence), is specifically designed to overcome existing barriers in the EO sector.
SPAI Innovations
• User-Friendly Interface for Non-Experts: Utilizes Natural Language Processing (NLP) to enable non-experts to easily extract actionable information from EO data without requiring specialized knowledge.
• Advanced Interface for Experts: Offers a Domain Specific Language that allows data scientists to interact with EO data and AI models in near real-time, enhancing data manipulation capabilities.
• Cloud Optimized GeoTIFFs (COGs): Facilitates efficient, affordable cloud workflows, enabling seamless data storage, access, and processing on cloud platforms.
SPAI Procedural Innovations
1. Simplified EO Data Access and Processing: SPAI simplifies the handling of EO data, removing traditional barriers to access.
2. AI-Ready Data Preparation: Employs the SCAN labeling tool to create datasets specifically designed for training neural networks.
3. Open PytorchEO Library: Supports the design, training, and export of neural networks, making it easier for developers to work with AI models.
4. Model Repository: Provides a platform for the operational use of uploaded models, fostering collaboration and sharing within the community.
SPAI revolutionizes the EO processing chain, making it iterative and continuously learning, thereby reducing training costs and improving scalability.
Expected Impact
1. Proactive Disaster Management: SPAI enables adaptive responses to climate extremes, significantly reducing socio-economic damages by allowing timely interventions.
2. Enhanced Market Penetration and ROI: By increasing the uptake of EO data, SPAI enhances the ROI for investments in the space economy, encouraging further investment.
3. Increased Accessibility and Usability: SPAI democratizes satellite analytics, making it accessible to a broader range of users, including those without technical expertise.
4. Improved Scalability and Efficiency: SPAI enhances the scalability and cost-effectiveness of EO data processing, making it suitable for a wide range of applications and industries.
Overall, EarthPulse aims to transform the EO landscape by making satellite data analysis more accessible, efficient, and impactful, addressing critical climate change challenges and market opportunities.
SPAI underwent significant architectural improvements in the second half of 2023 to address limitations in flexibility and modularity. This pivot, completed in Q1 2024, resulted in a fully flexible geospatial intelligence platform that serves as a single point of access for satellite analytics. This new architecture allows us to develop services and pilot projects 50% faster than before and ensures seamless integration with digital products and software such as GIS tools, data platforms like Databricks, and intelligence platforms like Power BI.
Following the product pivot, comprehensive product validation was conducted through user interviews, confirming a strong product-market fit. The solution addresses key pain points such as the lack of access to geospatial data and the need for a user-friendly platform. Continuous validation processes are underway to further refine the product based on user feedback.
Several pilot projects and initiatives have expanded EarthPulse's reach into new fields, including deforestation advisory services, air quality assessment, and climate risk analysis. These projects have facilitated the development of cutting-edge algorithms and advanced the technical and scientific capabilities of the SPAI platform.
For instance, pilots in deforestation analysis have enabled the implementation of state-of-the-art algorithms to monitor and assess forest loss, aligned with the EU Deforestation Regulation (EUDR23). In air quality assessment, SPAI has integrated satellite data on atmospheric composition and developed models for pollutant dispersion and IoT data integration, enhancing the resolution and accuracy of air quality monitoring.
A pilot for the insurance sector developed a solution for assessing climate risk, incorporating new data sources such as digital terrain models and climate data, enriching SPAI's data capabilities. Another pilot in landslide risk assessment combined satellite, topographic, and meteorological data to evaluate slope stability, resulting in a comprehensive risk assessment tool.
These advancements underscore EarthPulse's vision to become the central hub for satellite analytics, regardless of data type or sector.
Following the product pivot, comprehensive product validation was conducted through user interviews, confirming a strong product-market fit. The solution addresses key pain points such as the lack of access to geospatial data and the need for a user-friendly platform. Continuous validation processes are underway to further refine the product based on user feedback.
Several pilot projects and initiatives have expanded EarthPulse's reach into new fields, including deforestation advisory services, air quality assessment, and climate risk analysis. These projects have facilitated the development of cutting-edge algorithms and advanced the technical and scientific capabilities of the SPAI platform.
For instance, pilots in deforestation analysis have enabled the implementation of state-of-the-art algorithms to monitor and assess forest loss, aligned with the EU Deforestation Regulation (EUDR23). In air quality assessment, SPAI has integrated satellite data on atmospheric composition and developed models for pollutant dispersion and IoT data integration, enhancing the resolution and accuracy of air quality monitoring.
A pilot for the insurance sector developed a solution for assessing climate risk, incorporating new data sources such as digital terrain models and climate data, enriching SPAI's data capabilities. Another pilot in landslide risk assessment combined satellite, topographic, and meteorological data to evaluate slope stability, resulting in a comprehensive risk assessment tool.
These advancements underscore EarthPulse's vision to become the central hub for satellite analytics, regardless of data type or sector.
EarthPulse's activities have deepened sectoral knowledge and opened new avenues for service provision. The project has identified key areas for improvement, such as the need for regulatory support to validate EO data certification processes and ensure compliance with EUDR23 regulations.
In the insurance market, understanding workflows and data usage has been key, but more pilots and demonstrations are needed to drive digital transformation within the sector. Continuous algorithm iteration is essential to leverage the latest scientific advancements and data availability.
In the insurance market, understanding workflows and data usage has been key, but more pilots and demonstrations are needed to drive digital transformation within the sector. Continuous algorithm iteration is essential to leverage the latest scientific advancements and data availability.