Abstract. As urbanization intensifies globally, the Urban Heat
Island (UHI) effect has emerged as a critical environmental challenge,
inducing higher energy demands, compromised air quality, and significant
public health risks. Vertical Greenery Systems (VGS) such as green
facades and living walls, offer a spatially efficient adaptation
strategy by utilizing vertical surface areas of the built environment
for thermal regulation and microclimatic improvement, yet the absence of
data-integrative methods hinders large-scale evaluation of factors
defining a surface’s suitability for VGS. Previous methodologies for
estimating city-wide greening potential have successfully integrated
semantic 3D city models (LoD2) with Street View Imagery (SVI) to derive
key suitability factors such as Window-to-Wall Ratio (WWR) and Solid
Wall Area (SWA). However, these approaches are inherently limited by the
sparse spatial coverage of SVI, which is restricted to navigable road
networks, leaving rear facades and inner courtyards unassessed, and
which is frequently obstructed by foreground occlusion, leading to
errors in the factor calculation. This consecutive work introduces a
robust computational method that enhances the existing LoD2-SVI pipeline
with Oblique Aerial Imagery extending the potential estimation to over
90% of the urban building stock. To mitigate occlusion and resolution
disparities, we propose a multi-view fusion algorithm that aggregates
detections across multiple views within one perspective and further
across two perspectives. Our evaluation demonstrates that both data
sources deliver comparable results when assessing identical facades.
Further, our fusion approach significantly reduces systematic biases
found in single-source estimations. Ultimately, while the fusion
approach maximizes assessment reliability for walls with dual coverage,
the integration of oblique imagery remains critical for scalability.
Although it yields lower feature fidelity than street view, it provides
the only viable means to assess surfaces lying beyond the navigable road
network. Reproducibility review available at: https://doi.org/10.17605/OSF.IO/HETV8
