Linking photography behaviour and spatial composition in Suzhou’s classical gardens through computational geotagged image analysis
Understanding how visitors photograph classical gardens can offer new insights into the interaction between spatial composition and visitor engagement. Using the Classical Gardens of Suzhou as case studies, this study integrated UGC geotagged photographs with computer vision-based spatial analysis to examine this relationship. This study mapped the spatial distribution of photography behavior and identified hotspots, quantified scene characteristics through semantic segmentation and depth estimation, compared photography density across spatial composition types, and examined these differences in terms of overall spatial features and layered element organization. Four spatial composition types were identified. Differences in photography density were interpreted through the overall coordinated organization of element diversity, dominance, openness, and depth variation, while foreground–midground–background element distributions clarified how each type was formed. This study offers a data-driven approach for linking geotagged photography behavior with spatial composition and provides evidence for heritage promotion and dissemination in the digital era.
Authors
- Cheng Wei (ORCID: https://orcid.org/0009-0001-5971-3887)
- Yijing Wang (ORCID: https://orcid.org/0000-0002-5306-4994)
- Jing Jia
- Yujie Du
- Xinru Huang
Institutions
- Nanjing Tech University (CN)
- Tarim University (CN)
Publication Details
- Journal
- npj Heritage Science
- Published
- 2026-11-21
- DOI
- https://doi.org/10.1038/s40494-026-03073-7
- Primary Topic
- Urban Design and Spatial Analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00