Landscape

[E1-01] Capturing Walking-related Perceptions and Willingness within Tokyo's Station Areas: Leveraging Crowd-sourced Methods and AI Approach

*LU HUANG¹, TAKUYA OKI¹, SACHIO MUTO², YOSHIKI OGAWA² (1. Tokyo Institute of Technology, 2. The University of Tokyo)

Keywords:walking, perceptions, willingness, crowd-sourced method, AI

Walking in Tokyo, particularly around rail transit stations, plays a critical role given the city's high population density, compact urban structure, and heavy reliance on public transportation. With a growing recognition of the subjective aspects of walking, including perceptions and willingness, this study utilizes street view big data and AI technologies to delve into these aspects. Firstly, we gather an extensive collection of public perceptions and willingness through crowd-sourced image comparison questionnaires. Subsequently, by applying deep learning techniques, we achieve a milestone by training a model capable of automating the measurement of nine dimensions associated with walking. This accomplishment not only lays the foundation for automating subjective pedestrian assessment but also holds promise in supporting street design practices and facilitating further research to enhance walkability at the street level.