Exploring the impact of urbanization intensity on human perception through street view imagery

Conventional remote sensing–based measures of urbanization characterize built–unbuilt patterns at coarse spatial scales, limiting their ability to capture localized, human-scale features that shape how cities are experienced. To address this gap, we propose a street-view–derived Urbanization Intensity Index (UII) that operationalizes urbanization as a dual-channel construct combining built-environment exposure and streetscape functional diversity. Using New York, Hong Kong, and Mumbai, we compare UII with three remote-sensing indicators (ISA, UDL, LUI) and relate them to six perceptions (safety, liveliness, wealth, beauty, boredom, depression) predicted by a validated SVI model. Across cities, UII correlates positively with liveliness, safety, and wealth and negatively with boredom, while beauty exhibits context-dependent trade-offs in denser settings. Nonlinear fits further reveal threshold-like regimes: perception scores change most strongly as UII rises through ≈0.2–0.7, whereas beyond a risk zone (often >0.7) the curves flatten and can become adverse (downturns in safety/liveliness/wealth and rebounds in boredom/depression). Cross-validated calibration against human judgments shows UII aligns more closely with perceived urbanization than conventional overhead indicators. Spatial analyses indicate that UII hotspots co-locate with lively/wealthy/safe clusters—and cold spots with boring/depressing clusters—at finer corridor scales than remote-sensing proxies. Overall, these findings highlight the dual role of urbanization in covarying with positive and negative perception dimensions in city-specific ways, and underscore the need for planning strategies that balance development exposure with streetscape diversity, visual openness, and natural elements. • Develop a novel Urbanization Intensity Index (UII) from street view imagery to measure built–natural balance and structural diversity at the human scale. • UII correlates positively with liveliness, safety, and wealth and negatively with boredom, while beauty shows city-specific trade-offs in denser contexts. Cross-validated calibration against human ratings indicates that UII is better aligned with perceived urbanization than conventional indicators. • Identify the fine-grained spatial distribution patterns in the relationship between urbanization intensity and perception, demonstrating superior sensitivity in detecting local variations and built environment nuances.