Tule mukaan kuulemaan, kuinka tekoäly vauhdittaa satelliittitietojen käsittelyä Tanskassa! Entä miten tekoälyä sovelletaan sorateiden ja polkujen kartoittamisessa Islannissa? Mukana jo yli 120 osallistujaa!
Kuva: GI Nordenin tekoälytapahtuma järjestetään virtuaalisesti klo 15:00 Suomen aikaa.
Rekisteröidy mukaan Tanskan GeoForumin sivuilla. Löydät rekisteröitymislomakkeen sivun alalaidasta! Yleisö voi osallistua keskusteluun chatin kautta.
15:00 Welcome to the GI Norden webinar. Sofi Almqvist, Geoforum Sweden
15:05 Why AI is the next step in digitalisation. Mariell Juhlin, CEO of consulting company Policy Impact AB, Sweden
As an experienced economist and expert advisor, Mariell Juhlin will clarify why AI is the next step in digitalisation. She will give a status of the situation today and what needs to change to achieve the untapped societal benefit of AI. What should we do to make a difference, and what are the lessons learnt and recommendations thus far? Mariell will illustrate her points using example use cases from the implementation and use of AI within city planning, construction and real estate management based on data-driven strategy, innovation and business development.
15:20 The trend of AI in mapping: from 2D to 3D. Dr. Lingli Zhu, National Land Survey of Finland
As the project leader of the AI project, Advanced Technology for National Topographic Map updating (ATMU), Lingli Zhu will do a presentation to describe the AI development, state of the art, and its trend in the field of mapping, as well as the exploration of deep learning in 2D and 3D in the ATMU project in National Land Survey of Finland. With a Doctoral Degree in the fields of Photogrammetry and Remote Sensing, from Aalto University, Finland, and with numerous scientific papers and educational books from her hand, she is a real expert in the fields of remote sensing, sensors, geoinformation, 3D modelling, and virtual reality.
15:35 On the use of geospatial AI in Norway. Alexander Nossum, Norkart, Norway
Norway has a long tradition in public-public and private-public collaboration resulting in massive amounts of high detailed aerial photos with often year-to-year updates and centimeter resolution. Combined with national standards on map data and rich cadastre models this serves as a perfect point of departure for innovative use of novel AI methods. Alexander holds a ph.d. in geomatics, has over ten years experience working in the intersection between private-public collaboration and innovation in one of Norway’s largest geospatial IT companies, Norkart. In this presentation he will highlight the “KartAi ecosystem of projects” as well as other relevant geospatial AI projects from both the public, the private and the R&D sector. The KartAi ecosystem of projects revolves around the idea of harvesting the potential of AI on aerial photos combined with citizen participation in order to more efficiently and in an automated way to obtain an accurate cadastre and map data representation of the built environment.
15:50 The combined power of AI and satellite data to deliver scalable and dynamic monitoring solutions. Mads Christensen, DHI, Denmark
Earth observation satellites are our modern worlds ‘macroscope’, providing a critical global-scale view of earth and its processes. In fact, every day, several hundreds of terrabytes of new satellite imagery is collected worldwide. While such vast amounts of data provide unprecedented opportunities to track and monitor environmental change in near real time – from local to global levels – most of this data never meets the human eye. The fact is that we are not enough people on Earth to examine the daily amount of satellite data generated. Consequently, AI-based approaches have become vital instruments for processing and analyzing satellite imagery. In this presentation, we will illustrate how we leverage AI to improve and to accelerate satellite data processing, and how this can be used to underpin new and scalable monitoring solutions in many different domains.
16:05 Road extraction from aerial images using deep learning in West Iceland. Asra Salimi, Svarmi, Iceland
Nowadays, automatically extracting roads from satellite or aerial images can be necessary and helpful in many ways. For example, updating maps, urban planning, automotive navigation, and emergency and rescue systems can be mentioned as some of its applications. On the other hand, there are a lot of gravel roads and offroad paths (4×4 paths) that are very rough and unclear, made by local people which are not on the general road map of Iceland. Mapping these roads is very challenging, time-consuming, and manpower requiring. We recently developed a fully automatic method to extract gravel roads (rough 4×4 paths) and paved roads in the West of Iceland using deep learning. The deep learning algorithm has been trained and tested using a publicly available drone data along with 130TB drone data of our own.
16:30 Webinar ends
Rekisteröidy mukaan Tanskan GeoForumin sivuilla. Löydät rekisteröitymislomakkeen sivun alalaidasta!