Combined effects of the visual-thermal environment on restorative benefits in hot outdoor public spaces: A case study in Shenzhen, China
DONG Wen, DAI Donghui, LIU Mei, WANG Yaowu, LI Shuang, SHEN Pengyuan
2025
Building and Environment

Semantic segmentation and deep value computation based on deep learning.
Summary
This study explores how visual-thermal environments in high-density cities impact well-being through deep learning and field measurements. Findings reveal solar radiation dominates restoration in low green view areas (GVI <30%), while high spatial order mitigates thermal discomfort (>600W/m²). Landscape depth improves restoration in high-GVI spaces under low wind (<1.5m/s), offering integrated design strategies to enhance urban health amid global warming and urbanization.
Abstract
In the context of global warming and rapid urbanisation, improving visual-thermal environments in public spaces is key to enhancing well-being in high-density cities. This study utilizes deep learning techniques and ffeld measurements to quantify visual and thermal environment factors. It examines the contribution of inffuencing factors in different green view spaces and reveals how to modulate multisensory experience through visual factors under speciffc thermal environments to promote recovery beneffts. The result shows that thermal factors play a signiffcant role, with solar radiation being the most important factor affecting restoration in low green view index spaces (GVI < 30 %). Furthermore, the study revealed that high orderliness could alleviate thermal discomfort (solar radiation > 600W/m²) and promote overall restoration. Additionally, under calm or light air conditions (wind speed < 1.5 m/s), enhancing landscape depth can facilitate restoration in spaces with high GVI. Our research allows for a deeper understanding of the potential value of the combined effects of visual-thermal environments in enhancing residents’ health under intensiffed heat exposure. It also provides important impli-cations
Article Content
Typing...
The full content is currently under development.

Semantic segmentation and deep value computation based on deep learning.
Publication Details
Journal
Building and Environment
Publication Year
2025
Authors
DONG Wen, DAI Donghui, LIU Mei, WANG Yaowu, LI Shuang, SHEN Pengyuan
Categories
Synergizing comfort and energy efficiency in the built environment