Perceptual characteristics of urban street landscapes in Gaoling District based on street view image data

The perceptual characteristics of urban streets are crucial for urban planning and renewal, yet existing research often focuses on singular landscape elements and lacks multidimensional analysis. This study takes five streets in the Gaoling District of Xi’an as examples, utilizing Baidu Street View and semantic segmentation techniques to analyze street perceptual features from five dimensions: greenness, openness, enclosure, walkability, and imageability, while also exploring their spatial heterogeneity. By constructing perceptual vectors and applying K-Means clustering, streets with similar perceptual characteristics were categorized, providing new perspectives for urban street planning. The clustering results indicate the following: (1) The perceptual characteristics of the five sampled streets in Gaoling District exhibit significant spatial heterogeneity, with greenness showing a distribution pattern of more in the south and less in the north, and east more than west. Openness and enclosure display inverse distributions, with openness being stronger in the north and weaker in the south, while enclosure shows the opposite trend. The imageability of the sampled streets demonstrates a clustered distribution, with perceptual features being evenly distributed. (2) Based on perceptual vector characteristics, all street sample points were classified into three categories, revealing new types. Spatial differences were small in openness, imageability, and greenness, while walkability and enclosure exhibited more significant spatial disparities. Category C1 streets have stronger enclosure and walkability, while other perceptual features showed smaller differences. Category C2 streets had relatively balanced perceptual characteristics, and Category C3 streets exhibited the strongest enclosure, with smaller differences in other perceptual features. This classification enhances the understanding of the spatial heterogeneity of urban street landscapes and provides important insights for urban planning and renewal strategies.