KIT | KIT-Bibliothek | Impressum | Datenschutz

AI-based thermal bridge detection of building rooftops on district scale using aerial images

Mayer, Zoe; Kahn, James ORCID iD icon; Hou, Yu; Volk, Rebekka ORCID iD icon


Thermal bridges are weak areas of building envelopes that conduct more heat to the
outside than surrounding envelope areas. They lead to increased energy consumption and the
formation of mold. With a neural network approach, we demonstrate a method of automatically
detecting thermal bridges on building rooftops from panorama drone images of whole city
districts. To train the neural network, we created a dataset including 917 images and 6895
annotations. The images in the dataset contain thermal information for detecting thermal bridges
and a height map for rooftop recognition in addition to regular RGB information. Due to the small
dataset, our approach currently only has an average recall of 9.4% @IoU:0.5-0.95 (14.4% for large
objects). Nevertheless, our approach reliably detects structures only on rooftops and not on other
parts of buildings, without any additional segmentation effort of building parts.

Verlagsausgabe §
DOI: 10.5445/IR/1000136256
Veröffentlicht am 09.08.2021
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Industriebetriebslehre und Industrielle Produktion (IIP)
Steinbuch Centre for Computing (SCC)
Universität Karlsruhe (TH) – Zentrale Einrichtungen (Zentrale Einrichtungen)
Publikationstyp Proceedingsbeitrag
Publikationsjahr 2021
Sprache Englisch
Identifikator ISBN: 978-3-7983-3212-6
KITopen-ID: 1000136256
HGF-Programm 46.21.04 (POF IV, LK 01) HAICU
Erschienen in EG-ICE 2021 Workshop on Intelligent Computing in Engineering. Ed.: Jimmy Abualdenien, André Borrmann, Lucian-Constantin Ungureanu, Timo Hartmann
Veranstaltung 28th EG-ICE International Workshop on Intelligent Computing in Engineering (2021), Berlin, Deutschland, 30.06.2021 – 02.07.2021
Auflage 1
Verlag Universitätsverlag der TU Berlin
Seiten 497–507
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page