Speaker
Description
In recent years, street spaces have faced governance challenges characterized by dual disorders in both physical environment and social values, with the transition from traditional "growth-oriented" to "equity-oriented" development still encountering numerous obstacles. Existing studies have lacked systematic evaluation tools and tended to emphasize physical environment improvement while neglecting spatial equity, making it difficult to support refined governance decisions. Based on spatial justice theory, this study integrated street view imagery and deep learning technology to construct a coupling analysis framework for the evolution of street landscape perception and publicness transformation in Shanghai, China, exploring their dynamic evolutionary coupling patterns. The findings revealed that: (1) While street landscape perception showed an overall improving trend with significant spatial differentiation, with Shanghai's core areas displaying radial improvement patterns, publicness exhibited greater volatility, with central areas (particularly waterfront districts) even showing declining trends; (2) Five typical coupling modes were identified: "Synergistic Improvement" "Imbalanced Evolution Type I" "Imbalanced Evolution Type II" "Dual Recession" and "Stable Development" with 38.94% of streets found to be in non-benign evolutionary states, each type demonstrating differentiated spatial clustering characteristics; (3) Multi-level analysis across location-function-policy dimensions revealed that Shanghai's central urban areas tended to prioritize landscape improvement over publicness, while historic preservation districts faced challenges in maintaining public attributes. This study broke through the static research paradigm by constructing an AI-based street space evaluation system from a dynamic evolutionary perspective, providing theoretical foundation and practical pathways for advancing spatial renewal from mere aesthetic enhancement toward deeper social equity considerations in urban development worldwide.
| Keywords | Street Landscape; Publicness; Coupled Evolution; Spatial Justice; Deep Learning |
|---|---|
| Best Congress Paper Award | Yes |
