Research on Influencing Factors of Visual Safety Perception of Cycling Environment in High-density Residential Areas:A Case Study Based on Wuhan

ZHANG Yuanxin; WANG Luqi; XU Tao

doi:10.3969/j.issn.1000-0232.2025.03.002

Abstract：Cycling plays a vital role in reducing carbon emissions from urban transportation and improving residents'physical and mental well-being.A safe cycling environment is a fundamental prerequisite to encourage residents to choose green travel.It requires a deep understanding of influencing factors of cycling safety driving to improve the cycling environment's safety.High-density residential areas are important places for cycling activities in urban areas.In this study,three typical high-density residential areas in the main urban area of Wuhan City were chosen as the research samples.Multiple mixed-effect ordinal models were built by the self-collected cycling street view data and questionnaire surveys.The influencing factors of visual cycling safety perception and their action mechanisms were analyzed from the perspectives of the environment and individuals. First,the cycling routes were designed within the selected districts,and a GoPro camera was mounted on a bicycle to record street-view videos from the cyclist's perspective.One image was extracted every 20 m using the GPS trajectory embedded in the videos.Computer vision technology was employed for panoptic segmentation of all extracted images.The environmental elements were identified,and indexes like a green view index and enclosure degree were calculated.Second,100 representative images were selected according to the calculated results of indicators and the types and colors of cycle paths for the questionnaire survey of cycling environmental safety perception.The scores of 588 respondents to cycling safety perception in images and their personal information(e.g.,age,gender,and daily cycling behaviors)were collected in the questionnaire.A joint analysis of environmental data in images and questionnaire data was carried out.Considering the nested nature of data,some mixed-effect ordinal models were built to analyze the influences of individual and environmental factors on cycling safety perception.Additionally,models were built for five types of cycling scenarios(intersections,roads without dedicated cycle path,cycle path separated by road markings,cycle path separated by guardrails,and cycle path separated by isolation stripes),aiming to analyze differences in safety perception in different environments. Results showed that safety perception varies significantly across different types of cycling scenarios.Both environmental and individual factors have important influences on cycling safety perception,and effects of the same factor vary to some extent across different scenarios.Given the sound construction of cycle paths,cycling safety perception is mainly affected by environmental factors.Green view index,sky openness,and enclosure degree have significantly positive effects on improved safety perception,whereas motor vehicles have significantly negative effects.Given a poor construction of cycle paths,individual factors become more dominant.Women and more experienced cyclists are more likely to feel unsafe.The type of cycle paths is the most significant factor influencing cycle safety perception.If the separation level between the cycle path and vehicle path is higher,the cycling safety perception is higher.However,guardrails and isolation stripes have a similar effect on enhancing cycling safety perception.Among the five types of cycling environments,the cycle path separated by road markings is influenced by more environmental factors.At intersections,the colored cycle lanes and interface diversity have positive effects on safety perception.Based on research results,some strategies to improve the safety of different types of cycling environments are proposed.Research results can provide some empirical references to constructing cycling environments and improving safety in high-density residential areas.