Classification of Urban Roads from a Near-Person Perspective:An Objective Measurement Based on Street View Elements
In order to optimize the urban road classification system and make up for the problem that the traditional classification system is constructed by car-based and ignores the human perspective,a data-driven method of classifying urban road types from the near-person perspective is proposed with the support of computer deep learning technology and open-source data.Taking Dalian city as a practical case,we extract and cluster the main components of streetscape elements,classify urban roads according to the combination of streetscape in road sections,and analyze the classification results in terms of image element composition,spatial distribution,quantitative structure,and surrounding land use,so as to realize the classification and interpretation of urban roads in Dalian city from a near-person perspective.
万方数据
