Walkability measurement of street space in Hefei based on streetscape data

Keywords: Streetscape Data; Walkability; Street Space; Hefei City
Abstract Type: Virtual Lightning Paper Abstract

Abstract

Abstract: Walkability of urban street space is an important manifestation of urban vitality, and scientific exploration of street space walkability indexes and techniques is of great significance for fine urban governance. This paper uses deep learning and streetscape image technology to optimize the WalkScore evaluation idea, simplify its calculation method, and add Baidu streetscape image data to build a street space element identification and walkability evaluation method through semantic segmentation software based on deep learning, taking three central districts of Hefei city as research objects. The results showed that the average green view rate and walkability of streets in the three central districts of Hefei were 0.353,36.724 in Old City, 0.348,30.007 in Government Affairs and 0.352,31.037 in Binhu New District, with obvious spatial differences. The distribution characteristics of walkability are "high inside and outside, low in the middle". The results show that the green view rate of streets is mainly influenced by the age of street construction, street function and street properties, while the walkability is mainly influenced by the service level of facilities and the spatial layout of daily service facilities. The research results are important for guiding the construction of a more reasonable street space in Hefei.