Experimental statistics Remote sensing data and artificial intelligence in the register census


27 May 2022 - As the concept of the register census is being designed, a feasibility study investigates in how far remote sensing data such as satellite data or aerial photographs can help to produce census results in the future. It is planned to completely cover the characteristics of buildings and dwellings on a register basis from 2031. To serve as a data basis, a register of buildings and dwellings will be set up as an administrative register.

The register census collects data on buildings with residential space, temporary accommodation (simple structures such as portacabins or hunting cabins) and the dwellings they contain. The register of buildings and dwellings is to include also buildings without residential space. Other countries’ experience shows that supplementary quality assurance measures can support the collection of complete and up-to-date information on buildings and dwellings. Temporary accommodation, for example, may be missing in the survey and can thus be included.

Under an innovation project of the Federation entitled “Sat4GWR_IF-Bund – Remote sensing & AI for the register census“, algorithms will be developed on the basis of neural networks by means of which buildings can be identified on satellite images or aereal photographs. Then specific characteristics can be derived such as the use and size of the building. Remote sensing data are to support quality assurance as the data of buildings in the register of buildings and dwellings are checked and processed. This complements the establishment of the register and its maintenance.

In this project, the Federal Statistical Office is collaborating closely with the German Aerospace Center and the Federal Agency for Cartography and Geodesy. The project was launched in October 2021 and is envisaged to run until December 2024. The results will then be integrated into the register census and tested until 2028.


More information on the topic from official statistics