| Title |
Analysis of Visual Characteristics of Urban Street Elements on Walking Satisfaction in Seoul, Korea - Application of Google Street View and Deep Learning Technique of Semantic Segmentation |
| Authors |
박근덕(Park, Keundeok) ; 기동환(Ki, Donghwan) ; 이수기(Lee, Sugie) |
| DOI |
https://doi.org/10.38195/judik.2021.06.22.3.55 |
| Keywords |
가로환경; 보행만족도; 구글 가로이미지; 컴퓨터 비전; 딥러닝; 의미론적 분할 Street Environment; Walking Satisfaction; Google Street View; Computer Vision; Deep Learning; Semantic Segmentation |
| Abstract |
This study focuses on the quantification of urban street elements, and the analysis of the relationship between urban street elements and walking satisfaction. The result of the study shows that quantified street elements have significant relationships with walking satisfaction. First, buildings, automobiles, and signboards have negative associations with walking satisfaction due to the closedness and lack of aesthetic impression. This finding indicates that the walking environment can be improved through the separation of pedestrians from vehicles and the maintenance of complex signboards. Second, street trees (e.g., trees, shrubs, flowers) have positive associations with walking satisfaction. This finding suggests that walking satisfaction can be improved through the implementation of greenery strategies for roadways and buildings. Third, this study finds that visual complexity has an inverted U-shaped relationship with walking satisfaction. In other words, a monotonous street environment or an excessively complex street environment has a negative association with walking satisfaction, suggesting that an appropriate level of visual complexity is important for walking satisfaction. |