Landscape elements and mental restoration in virtual environment: based on image analysis

Abstract

Modern landscape architects and researchers are actively discussing the role of landscapes as a means to contribute to mental health. However, despite these efforts, it remains unclear which landscape elements within the scene directly affect mental restoration. This study employs image analysis to detect and categorize landscape elements within physical environments. Based on this analysis, a virtual environment is created to simulate realistic landscapes and manipulate the elements to generate diverse scenes. Participants experience these virtual landscape scenes and their restoration degree was measured using scales for mental restoration and perceived sensory dimensions. Through this study, we aim to identify the landscape elements within the scene that directly influence mental restoration and determine the ideal proportions of these elements for restoration. Additionally, we provide complementary insights into perceived sensory dimensions that can further enhance mental restoration. This research contributes both scientifically, in terms of landscape scene analysis and virtual landscape generation methodologies, and practically, by informing the design of specialized spaces for mental restoration within landscape architecture.