A Walking Environment Assessment Based on Multisource Data-A Case Study of Ping'an Street in Beijing

Street oriented walking environment assessments are of great significance to the development of green transportation. 35 groups of eye movement data of pedestrians walking along the street were collected based on the eye movement instrument. The factors were identified and detected by using segnet and Yolo image recognition algorithms. The influencing factors of walking environment were screened based on physiological indicators, and the walking and riding environment index system for convenience, safety and comfort was established to construct the street scale walking index. Using the renewal of Ping'an Street in Beijing as an example, an empirical study was carried out on the 16km roads on both sides of Ping'an Street before and after the renewal, and the correlation analysis was carried out in combination with the residents' travel survey data of 17 surrounding residential areas. We found that after street renewal, the walking environment of Ping'an Street in Beijing is improved. With the renewal of the street, the walking of the street shows a decreasing trend from the center to the periphery, in which the decreasing trend of the convenience index is the most obvious, followed by the safety index. There is a significant positive correlation between walking index and walking travel sharing rate of community residents. Finally, we propose policy suggestions for the improvement of walking environment in Beijing.