Introducing a Perceptual–Spatial Landscape Planning Model (PSLPM) for cultural landscapes’ route optimization: The case of Chengde Mountain Resort

A novel Perceptual–Spatial Landscape Planning Model (PSLPM) demonstrates significant potential for optimizing cultural landscape routes by integrating public perception with spatial characteristics. Using Chengde Mountain Resort as a case study, researchers analyzed 144 scenes and 43,879 social media comments to generate perceptual scores via few-shot learning and large language models. These scores were then coupled with heritage spatial data using advanced graph neural networks, leading to optimized routes that outperformed conventional methods in terms of spatial diversity, perceptual coherence, and coverage. The model avoids overconcentration in popular areas while enhancing interpretive richness across varied spaces, offering a transferable framework for balancing public perception and heritage preservation in urban cultural landscapes.

Lian, J., Nijhuis, S., Bracken, G., Bai, N., & Zhang, H. (2026). Introducing a Perceptual–Spatial Landscape Planning Model (PSLPM) for cultural landscapes’ route optimization: The case of Chengde Mountain Resort. Frontiers of Architectural Research, 15(5), 1572–1588. https://doi.org/10.1016/j.foar.2025.10.004