Visual Quality Evaluation of Urban Landscape Based on Image Analysis in Shenyang, China

学生名:孫 冬

研究テーマ: Visual Quality Evaluation of Urban Landscape Based on Image Analysis in Shenyang, China

(中国瀋陽市の都市景観における画像分析を用いた視覚的印象評価に関する研究)

入学年月:2019.04

修了年月:2022.03

取得学位:博士（工学）

論文概要:Urban landscape quality plays a critical role in enhancing urban living environments and addressing challenges posed by rapid urbanization. This research develops a multi-level visual quality evaluation framework for urban landscapes using Shenyang, China, as a case study. The study integrates macro, meso, and micro perspectives, employing advanced methods such as deep learning-based semantic segmentation, Scenic Beauty Estimation (SBE), and spatial statistical analysis.

The research focuses on evaluating streetscapes, waterfront parks, historical districts, and residential areas. In the core urban area, the streetscape evaluation combines semantic segmentation with SBE and virtual reality to construct a visual quality model, revealing how green space, sky view, and building proportions influence public perception. For waterfront parks, the study identifies critical factors such as natural features and human-made elements impacting aesthetic quality, providing actionable insights for urban planning. Historical and cultural districts are analyzed to understand the role of architectural heritage and greenery in enhancing visual quality, while residential area evaluations highlight opportunities to improve landscape elements like green spaces.

The findings contribute to the development of an urban visual environment impact evaluation system and offer practical strategies for urban regeneration, aligning with sustainable development goals. By leveraging multi-source data and advanced technologies, this research establishes a robust foundation for enhancing urban landscape planning and management.