Examining the Influence of Isovist Geometry and Visual Perception on Spatial Diversity along Street Canyons

Street canyons, spaces between parallel building rows, are fundamental components of the urban fabric. With the rising complexity of urban environments and the emphasis on sustainable development, the diversity of these linear spaces has garnered increasing attention from scholars. However, previous studies primarily focused on the environmental aspects of street canyons, and the role of geometric visibility and visual perception in shaping their spatial diversity remains underexplored. To address this gap, this study examines the impact of vision-related attributes on spatial diversity along streets by employing advanced analytical methods. Specifically, spatial diversity is quantified based on variation of building category and height. Geometric visibility is assessed through Isovist analysis, which captures the visible area and configuration from specific observation points, while visual perception from street view images is processed using image segmentation to classify and evaluate urban elements. Furthermore, machine learning models, including k-nearest neighbors (KNN) and random forests, are utilized to investigate the potential non-linear relationships between visibility-related metrics and street canyon diversity. This approach provides new insights into how visual complexity interacts with structural features to influence urban form at the street level, contributing to the broader discourse on urban morphology and sustainable city planning and design.