Year: 2025
Type: article
Abstract: 3D city maps play an essential role in various applications, including emergency services, environmental impact assessment, urban planning, and location-based service. Existing approaches for building height estimation largely rely on expensive technologies such as LiDAR or aerial photogrammetry. This paper proposes a novel approach for building height estimation leveraging ubiquitous cellular network signals. We collected cellular data using the BitToBrick app in six areas to measure buildings with different types and roof shapes. The collected received signal strength (RSS) and 2D footprint of the buildings are used to estimate building heights through unsupervised learning methods. The results show that five out of six buildings had errors under 2.39 m, with the smallest errorbeing 0.02 m and the largest error being 4.68 m. This is comparable with the accuracy of Great Britain’s national mapping agencies. Our method has several advantages: it uses existing cellular infrastructure without specialised equipment, relies on stationary antennas to reduce computational overhead, and employs a lightweightmobile app to access mobile signals for building height estimation. These benefits showcase its effectiveness and scalability for this purpose.
Source: Big Earth Data
Authors: Yu Wang, Anahid Basiri
Institution: University of Glasgow
Language: English
Subfield: Building and Construction
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,024
Funder: UK Research and Innovation
Awards: G3007901435, G3741620067
OpenAlex ID: W7083681207