Min Wu, Kaige Zheng, Junhong Chen, Jiaxin Zhang, M Y Li, S Y Wu
Ensuring equitable park access amid rapid urbanization and population ageing is essential. We propose an integrated framework that couples street-network impedance modelling with streetscape visual perception to assess elderly walkability to parks in central Fuzhou, China. A multi-source dataset-street networks, street view images, park polygons and demographic data-was compiled. Physical accessibility was calculated via weighted path impedance, whereas perceived accessibility was estimated by semantic segmentation of images and machine-learning calibration against expert scores. Combining the two yields a composite walkability index that exposes areas where dense networks coexist with poor visual environments and vice versa, delineating green corridors and underserved neighbourhoods. By fusing objective and subjective dimensions, the framework identifies priority zones for age-friendly design and scalable park-planning interventions, and indicates where network repairs should be coupled with streetscape upgrades to reduce elderly green-access inequities, especially in underserved neighbourhoods.