GI_Forum 2019, Volume 7, Issue 1 Journal for Geographic Information Science
|
Verlag der Österreichischen Akademie der Wissenschaften Austrian Academy of Sciences Press
A-1011 Wien, Dr. Ignaz Seipel-Platz 2
Tel. +43-1-515 81/DW 3420, Fax +43-1-515 81/DW 3400 https://verlag.oeaw.ac.at, e-mail: verlag@oeaw.ac.at |
|
DATUM, UNTERSCHRIFT / DATE, SIGNATURE
BANK AUSTRIA CREDITANSTALT, WIEN (IBAN AT04 1100 0006 2280 0100, BIC BKAUATWW), DEUTSCHE BANK MÜNCHEN (IBAN DE16 7007 0024 0238 8270 00, BIC DEUTDEDBMUC)
|
GI_Forum 2019, Volume 7, Issue 1 Journal for Geographic Information Science
ISSN 2308-1708 Online Edition ISBN 978-3-7001-8609-0 Online Edition
Lukas Beer
S. 119 - 133 doi:10.1553/giscience2019_01_s119 Verlag der Österreichischen Akademie der Wissenschaften doi:10.1553/giscience2019_01_s119
Abstract: 3D city models play an important role in multiple applications, but creating them still requires effort using various possible techniques. This paper proposes a new machine-learning-based framework for generating 3D city models. With the help of conditional Generative Adversarial Networks and single orthographic images, segmentation and height estimations of buildings are achieved. The height information per pixel and the building coordinates were generalized using a histogram for heights and the Douglas-Peucker algorithm. The framework was evaluated by using variations of the same dataset (for the city of Berlin) to show possible differences due to changes in the image size and representation of the heights. The evaluation reveals that it is possible to generate block models with a mean absolute height error of 5.53m per building, a mean absolute height error for the whole raster of 1.32m, and a Jaccard Index of 0.55 for the segmentation. While the proposed framework for generating LoD1 city models does not attain the accuracy of previous techniques, our work represents a step towards successfully using machine learning for the automatic generation of city models and building segmentation. Keywords: city models, generative adversarial networks, LoD1, segmentation Published Online: 2019/06/19 08:21:36 Object Identifier: 0xc1aa5576 0x003aba41 Rights:https://creativecommons.org/licenses/by-nd/4.0/
GI_Forum publishes high quality original research across the transdisciplinary field of Geographic Information Science (GIScience). The journal provides a platform for dialogue among GI-Scientists and educators, technologists and critical thinkers in an ongoing effort to advance the field and ultimately contribute to the creation of an informed GISociety. Submissions concentrate on innovation in education, science, methodology and technologies in the spatial domain and their role towards a more just, ethical and sustainable science and society. GI_Forum implements the policy of open access publication after a double-blind peer review process through a highly international team of seasoned scientists for quality assurance. Special emphasis is put on actively supporting young scientists through formative reviews of their submissions. Only English language contributions are published.
|
Verlag der Österreichischen Akademie der Wissenschaften Austrian Academy of Sciences Press
A-1011 Wien, Dr. Ignaz Seipel-Platz 2
Tel. +43-1-515 81/DW 3420, Fax +43-1-515 81/DW 3400 https://verlag.oeaw.ac.at, e-mail: verlag@oeaw.ac.at |