Practical example and current challenges
- Driven by the development of smart cities, digital twins and 3D models are gaining great im-portance. However, processes related to the development of 3D spatial content bundle many re-sources and are cost and resource intensive. Very often, there are not enough resources to gen-erate the necessary data in a timely manner and to keep updating it appropriately once gener-ated.
- For the application of AI, the preparation and labeling of training data is an important part of the process. The challenge is to evaluate whether the 3D data and surfaces for selected cities and regions are generated with the required accuracy and whether the updates are performed frequently enough. The quality of results needs to be validated. In addition, the incoming data are of different semantics and structure. Therefore, it is crucial to standardize the data structure and semantics in order to process the data in a comprehensible manner and ensure a consistent high quality.
- The use case overcomes these hurdles by automating processes to create and update 3D surfac-es and objects in high detail using artificial intelligence methods. The goal is to develop solu-tions for European cities to automatically create and update relevant 3D content. Open Data Science concepts and technologies will be used. To keep the time-to-market as short as possi-ble, software products available on the market will be used to prepare the training data and pro-cess the incoming geodata. These data are urgently needed to meet societal challenges such as climate change, CO2 savings and modern climate-neutral mobility.
What added value does the "Gaia-X project" offer?
- Gaia-X assists in finding a suitable federated IT infrastructure to process the extensive amount of incoming data and train the artificial intelligence algorithm.
- A significant amount of cloud resources are required to train the artificial intelligence algorithm. Gaia-X enables the secure use of artificial intelligence across platform boundaries and provides a certified environment of mutual trust. Gaia-X enables intellectual property protection of the tagged training data and algorithm. In addition, Gaia-X promotes new digital business models for result exploitation.
- Gaia-X enables a reduction in the cost of creating 3D content and increases its quality and relevance. Federated production processes using AI are thus enabled by Gaia-X.
Use Case Team
- Michael Mundt – Esri Deutschland GmbH