Image-based Visibility Analysis Replacing Line-of-Sight Simulation: An Urban Landmark Perspective

TL;DR

The paper challenges the limitations of Line-of-Sight (LoS) based visibility analyses by adopting an image-based approach that uses Street View Imagery (SVI) and Vision-Language Models (VLM) to detect landmark visibility and to build a heterogeneous visibility graph that encodes visual-spatial relationships. It introduces a three-step detection pipeline—relative positioning, zoomed-in SVI regions, and zero-shot landmark detection via a VLM (OWL-ViT)—and couples this with a graph framework capturing Inter-visibility, Visual Co-existence, and VAV paths, enabling analysis of both individual viewpoints and city-scale visual networks. In two case studies, the method achieves high accuracy ($0.87$) for landmark visibility and reveals distinct visual contexts: comparison with 3D simulations shows variable alignment across cities, while Flickr data highlights tourist-oriented perception; along the River Thames, bridges account for about $0.30$ of VAV paths, underscoring their role as visual corridors. The study offers a scalable, context-aware alternative to 3D visibility analyses, with practical implications for urban planning, heritage conservation, and computational social science, and opens avenues for broader application to named places and visual-object networks.

Abstract

Visibility analysis is one of the fundamental analytics methods in urban planning and landscape research, traditionally conducted through computational simulations based on the Line-of-Sight (LoS) principle. However, when assessing the visibility of named urban objects such as landmarks, geometric intersection alone fails to capture the contextual and perceptual dimensions of visibility as experienced in the real world. The study challenges the traditional LoS-based approaches by introducing a new, image-based visibility analysis method. Specifically, a Vision Language Model (VLM) is applied to detect the target object within a direction-zoomed Street View Image (SVI). Successful detection represents the object's visibility at the corresponding SVI location. Further, a heterogeneous visibility graph is constructed to address the complex interaction between observers and target objects. In the first case study, the method proves its reliability in detecting the visibility of six tall landmark constructions in global cities, with an overall accuracy of 87%. Furthermore, it reveals broader contextual differences when the landmarks are perceived and experienced. In the second case, the proposed visibility graph uncovers the form and strength of connections for multiple landmarks along the River Thames in London, as well as the places where these connections occur. Notably, bridges on the River Thames account for approximately 30% of total connections. Our method complements and enhances traditional LoS-based visibility analysis, and showcases the possibility of revealing the prevalent connection of any visual objects in the urban environment. It opens up new research perspectives for urban planning, heritage conservation, and computational social science.

Figures (15)

• Figure 1: A research framework for the study. Imagery: Google Street View, Wikimedia Commons, Tripadvisor.
• Figure 2: Workflow of locating and detecting the visibility of a distant landmark from SVI. Imagery: Google Street View, Wikimedia Commons.
• Figure 3: Visualisation of the landmark detection process. (a) Locating landmarks using bounding boxes. (b) Detecting the landmark within the zoomed-in region.
• Figure 4: Left: Different types included in the graph definition. Right: An SVI taken near the London Bridge, showing how SVI can represent multiple edge relations via a single image. Imagery: Google Street View.
• Figure 5: Left: Query images of different landmarks investigated in the second case study. Imagery: Wikimedia Commons. Right: The spatial distribution of the selected landmarks along the River Thames. Basemap: © OpenStreetMap contributors.