Research on Spatial Perception and Physical Features of Urban Streets based on Street View Big Data and Computer Vision Technology
学生名:吕 美
研究テーマ:Research on Spatial Perception and Physical Features of Urban Streets based on Street View Big Data and Computer Vision Technology
(街路景観のビッグデータとコンピューター視覚技術に基づく都市街路の空間知覚と物理的特徴に関する研究)
入学年月:2020.04
修了年月:2023.03
取得学位:博士（工学）
論文概要:
Based on streetscape big data and computer vision technology, the streets of typical coastal cities Qingdao and Fukuoka were selected as the site to explain the correlation between physical features and the perceptual features of urban coastal streets to evaluation of the quality of street space.
IN CAPTER1, the study has located the problem of spatial design quality of coastal urban streets accurately, reflect the urban landscape, provide the basis for subsequent large-scale and deeper theoretical research of urban streets.
IN CAPTER 2, based on the web of science database, this chapter analyzes the highly cited literature in related fields with Cite Space visual analysis tool.
IN CAPTER 3, main research methods and measurement in urban streets were introduced.
In CAPTER 4, the analysis will be carried out in the sections of overall description of sampling point data, sampling point analysis, data trend analysis, comparison of road segments, and preliminary discussion of each physical feature combined with the current status of the street.
In CAPTER 5, the study used machine learning semantic segmentation, GIS and Semantic difference (SD) methods to obtain the spatial data and perceptual evaluation of coastal streets in Qingdao and Meiji streets in Fukuoka. Each of the six perceptual features, imageability, enclosure, human scale, transparency, complexity and nature, was taken as dependent variables and the corresponding physical features was taken as independent variables.
In CAPTER 6, Qingdao Coastal Streets, Eiji Street and Moji Streets in Fukuoka were selected as the study sites. Get sample photos on Baidu Street View and Google Street View through Python. By analyzing each established scale with SD method, the concept and structure of the object could be quantitatively described.
THE CAPTER 7 is CONCLUSION AND PROSPECT. Summarizes the conclusions of each chapter and the optimization strategies of urban streets were proposed.