The aim of the project is to develop the technology that allows for the automatic acquisition of building outlines and automatic acquisition of information about their height. This technology is to be based on machine learning, in particular on artificial neural networks. Monoscopic satellite images and metadata associated with these images are to serve as input data to the technological process. Obtaining the outlines of buildings and their heights from this type of data is an innovative solution of the Applicant, unique on a global scale.

The advantage of this type of solution, compared to the solutions currently available on the market, will be a faster and cheaper way to obtain 2.5D building models. These models are widely used in various fields, for example: spatial planning, telecommunications, noise analysis, and aviation. Reducing the costs of obtaining such models will allow access for entities and applications for which the price of this type of product was a barrier.

The project will be implemented in the period from 01.07.2021 to 30.06.2023.
Project value: 1,807,375.00 PLN Contribution of European Funds: 1,386,175.00 PLN

The „Machine Learning and Monoscopic Satellite Imagery for Generating 2.5D Building Models” project has been successfully completed, and its innovative products and services are now part of the KPGeo portfolio.

Using AI-powered machine learning, this technology extracts building outlines and heights from monoscopic satellite images, offering a faster and more affordable solution compared to traditional methods. For 80-90% of the data, the process is fully automated, while the remaining requires manual refinement to ensure top accuracy.

Learn more and contact us for details.