In the heart of Fuzhou, China, a groundbreaking study led by Xiaowen Zhuang from the College of Landscape Architecture and Art at Fujian Agriculture and Forestry University is revolutionizing the way we evaluate and optimize university outdoor spaces. Published in the journal *Buildings* (translated to English as “Buildings”), this research leverages cutting-edge technology to transform the often subjective and labor-intensive process of assessing restorative quality in campus environments.
Traditionally, evaluating the restorative quality of outdoor spaces has relied heavily on subjective surveys and manual assessments, which are not only time-consuming but also lack scalability and objectivity. Zhuang’s study addresses these limitations by applying explainable machine learning to predict restorative quality from campus imagery, enabling large-scale, data-driven evaluation. “This approach allows us to capture complex nonlinear relationships that traditional methods might overlook,” Zhuang explains.
The study focused on Fujian Agriculture and Forestry University, where researchers extracted road network data and generated 297 coordinates at 50-meter intervals. From these coordinates, 1197 images were collected, and surveys were conducted to obtain restorative quality scores. The Mask2Former model was employed to extract landscape features, and decision tree algorithms (RF, XGBoost, GBR) were selected based on MAE, MSE, and EVS metrics. The combination of optimal algorithms and SHAP (SHapley Additive exPlanations) was used to predict restoration quality and identify key features.
One of the most compelling aspects of this research is its ability to identify the specific elements that contribute to the restorative quality of outdoor spaces. “Natural elements like vegetation and water positively affect psychological perception, while structural components like walls and fences have negative or nonlinear effects,” Zhuang notes. This insight is crucial for designing outdoor spaces that not only look appealing but also promote well-being and social interaction.
The study also used a multivariate linear regression model to identify features with significant statistical impact but lower feature importance ranking. Additionally, k-means clustering was employed to analyze heterogeneity in scores for three restoration indicators and five campus zones. These findings provide a foundation for creating high-quality outdoor environments with restorative and social functions.
The implications of this research extend beyond the academic world. In the energy sector, for instance, understanding how natural elements impact psychological well-being can inform the design of green spaces around energy facilities, enhancing the overall quality of life for nearby communities. “By optimizing outdoor spaces, we can create environments that are not only aesthetically pleasing but also contribute to the health and well-being of the people who use them,” Zhuang says.
This study is a significant step forward in the field of landscape optimization, offering a data-driven approach that is both scalable and objective. As we continue to explore the intersection of technology and environmental design, the insights gained from this research will undoubtedly shape future developments in creating restorative and socially functional outdoor spaces.