GeoAI on show at INTERGEO & WSBW, cities most in need of trees, & how confidence helps build better models

Welcome to GeoAI Horizons - your monthly look at the intersection of geospatial data & artificial intelligence!

In this edition of GeoAI Horizons, we examine how GeoAI is being promoted at recent events such as INTERGEO and WSBW, which cities need trees most to sequester the estimated CO2 they produce each year, and how AI confidence can be used to help build better detector models.

GeoAI takes center stage: A growing focus at industry events

The 2024 INTERGEO Expo and Conference and World Space Business Week (#WSBW, by Novaspace ) provided a window into the rapidly evolving intersection of geospatial technologies and the space industry. One of the key themes at WSBW was the shared ambition across Earth observation (EO), satellite communications, and navigation to leverage space for societal good. With over 10,000 satellites now in orbit, the importance of satellite data in addressing critical issues such as climate change, urban planning, and disaster response has never been more evident. INTERGEO, meanwhile, spotlighted the role of tools like LIDAR and drone technology in providing high-precision mapping, which is essential for tackling environmental challenges and building more sustainable infrastructure.

Both events highlighted the increasing demand for real-time data and advanced analytics. Space situational awareness, a growing concern at WSBW, is driving a focus on protecting space as a resource for future generations. In Europe, the market for smaller launchers is growing, reflecting efforts to reduce costs and improve access to space. Despite EO making up a small portion of overall satellite launches, it remains critical for developing advanced Earth intelligence. At INTERGEO, the emphasis was on the practical application of geospatial technologies, from point cloud analysis to aerial imagery, showcasing their growing role in sectors like energy, construction, and telecommunications.

One of the most striking takeaways from both events was the rise of GeoAI as a critical tool for the future. Companies across industries are increasingly turning to artificial intelligence to process geospatial data, helping to transform vast datasets into meaningful insights for decision-making. From environmental monitoring to infrastructure development, GeoAI enables faster, more scalable, and more precise solutions. As the space and geospatial industries continue to evolve, the integration of AI will be key to unlocking new capabilities and addressing complex global challenges.

Read Pierrick Poulenas' wrap-up of WSBW.

Mapping the urban jungle: Which global cities need trees the most?

Urban trees are vital for enhancing environmental health and tackling climate change. They provide numerous benefits, including carbon sequestration, improved air quality, and temperature regulation. A recent analysis focused on 26 global cities to pinpoint where additional tree planting is most urgently needed. Tokyo emerged as the city requiring the highest number of new trees, with a staggering need for over 2 billion. In contrast, Rotterdam showed the least demand for tree planting, underscoring the disparities in urban greenery needs across different regions.

This initiative highlights the significance of reforestation and innovative monitoring techniques, such as geospatial AI, which can help assess and prioritize urban greenery. By leveraging technology to map CO2 emissions and identify areas lacking sufficient tree cover, cities can develop targeted strategies to enhance their environments. Integrating geospatial data allows urban planners to visualize the potential impact of new trees, fostering biodiversity and promoting healthier living conditions for residents.

The findings emphasize the importance of incorporating environmental considerations into urban planning and policy-making. Cities equipped with this information can better address their unique challenges and work towards sustainable solutions. By prioritizing tree-planting initiatives, urban areas can significantly improve their resilience against climate change and enhance the overall quality of life.

Check out this blog article for the full league table and more cool animations, such as the one above.

How AI confidence helps build better detection models

The accuracy of AI models is paramount for developing effective detection systems, and Picterra 's recent enhancements significantly improve this aspect through advanced confidence attributes. Each detection polygon now includes a "confidence" score, indicating how certain the model is about its predictions. This score provides valuable context, helping users identify areas where the model excels and where it may require further training or data. For example, a confidence score of 90% indicates a high level of certainty, enabling more informed decision-making.

The confidence score is computed using the model's output, closely linked to creating uncertainty maps in accuracy areas. By concentrating on the inner 50% of a detection area, the system mitigates the impact of less reliable outer edges, resulting in more accurate assessments. This functionality is effective in both count mode and segmentation mode, although it is particularly insightful in count mode, where large polygons yield meaningful confidence values.

Additionally, Picterra offers enhanced visualization tools that allow users to view confidence scores across detection layers using a color gradient. This feature effectively highlights regions where the detector is less certain, guiding users on whether to collect more training data or refine detection boundaries. Future updates are set to introduce bulk filtering options based on confidence levels, enabling users to enhance model precision and overall accuracy.

Watch this GeoAI Bytes session to discover how insights into a detector's confidence level can help us iterate more effectively and build more accurate geoAI models.

Engage with the GeoAI community

With event season in full swing, there are a wealth of upcoming events we will be attending where you can engage with other like-minded space and GeoAI professionals, including;

• EUSPA User Consultation Platform 2024: October 8. EUSPA - EU Agency for the Space Programme ’s User Consultation Platform (UCP) is a systematic consultation involving business and industry leaders, entrepreneurs, service providers, innovators, and other EU Space user communities to express their needs and requirements, share best practices, and present case studies. Picterra will participate in the roundtable on environmental impact monitoring for mines at 14:15.
• Space Security Symposium: October 9. This event is part of the inaugural UK-Swiss Space Exchange, which brings together experts and stakeholders to discuss key issues and advancements in space security. Organized by the Swiss Business Hub and the Swiss Defence Attaché under the patronage of H.E. Markus Leitner, it highlights the importance of international collaboration in the rapidly evolving space sector.
• International Astronautical Congress (IAC) 2024: October 14-18. The 75th International Astronautical Congress theme will be “Responsible Space for Sustainability.” The intention is to highlight the importance of Space as an environment that must be kept secure and open to exploration, peaceful use, and international cooperation by present and future generations in the interests of the planet and all nations, regardless of their level of development and without discrimination of any kind.

That's it for this edition of GeoAI Horizons. Be sure to subscribe so you don't miss future editions! Do you have feedback or something you'd like us to cover? Don't hesitate to get in touch.