Visual perception-based analysis of community age-friendliness

This study develops a framework for evaluating community age-friendliness through street-view imagery, aiming to elucidate the relationship between environmental visual features and older adults’ perceptions of livability. Using Weifang City as a case study, a total of 172,000 Baidu street-view images (2024) were analyzed with the BEiT semantic segmentation model pretrained on ImageNet. Four key environmental indicators were extracted: Green View Index, Street Enclosure, Sky Openness, and Street Walkability. The results indicate that Street Walkability exhibits the strongest correlation with subjective assessments of age-friendliness (R² = 0.421), followed by Sky Openness (R² = 0.241), Green View Index (R² = 0.220), and Street Enclosure (R² = 0.103). Spatial autocorrelation analysis further reveals significant clustering patterns for Walkability and Sky Openness, with contrasting distributions between central and peripheral areas. These findings underscore the critical role of streetscape visual features in shaping perceptions of age-friendly communities and demonstrate the potential of integrating computer vision and remote sensing techniques into urban planning and policy for aging societies.