The Study on the Measurement and Enhancement Strategies of Urban Street Space Quality Based on Street View Images: A Case Study of Streets in the Central Urban Area of Xi’an

Abstract

Identifying the spatial quality characteristics of urban streets is crucial for achieving scientific planning and effective management. This provides a basis for optimizing urban spatial layouts and promotes the intelligent and sustainable development of cities. However, most previous studies have been limited to specific neighborhoods, lacking a systematic analysis of street quality across large urban areas. Therefore, this study integrates spatial pattern data and street landscape image data from the main urban area of Xi’an, constructing a quantitative database for street spatial quality. It establishes a measurement system from both objective physical space and subjective perception. The system covers four primary dimensions: safety, comfort, richness, and convenience, with fourteen evaluation indicators. The system uses the entropy value method to assign weights and completes a large-scale quantitative assessment of street spatial quality. The study results indicate: (1) The spatial quality of streets in the main urban area of Xi’an generally shows a ‘high in the center, low in the periphery’ layered distribution pattern. The central urban area has significant advantages in street interfaces, visual environments, and traffic organization, while the peripheral new districts face issues such as insufficient greenery, enclosed spaces, and single-function configurations. (2) Through K-means clustering, this study identifies four types of street spatial qualities: safety-oriented, ecological comfort, interface vitality, and pedestrian convenience. It further analyzes the causes and mechanisms of these types from four aspects: location conditions, land use, planning policies, and historical evolution. (3) The study also used the Cohen’s s Kappa coefficient to test the consistency between model outputs and subjective perception levels. The results showed that the integrated model outperformed the single objective index model in terms of recognition accuracy, confirming the necessity of incorporating perception dimensions. (4) Enhancing the spatial quality of streets requires a multi-dimensional collaborative strategy, including enriching street interfaces, enhancing road surveillance and lighting design, and improving green spaces and pedestrian facilities. The framework for measuring street spatial quality proposed in this paper is highly scalable and adaptable, both methodologically and practically. It not only provides a basis for targeted measures to improve the quality of streets in Xi’an’s main urban area but also offers a feasible path and technical support for other cities to advance the refined governance and human-centric optimization of public spaces.