Distribution and preference analysis of Streetscape visual elements based on street view images- Focused on Jung-gu, Gangnam-gu, Seocho-gu, and Songpa-gu in Seoul -

초록

Are urban streets beautiful? This question not only influences residents’ daily quality of life but is also a key factor in attracting tourists. To explore differences in beauty perception between residents and tourists, this study proposes a new evaluation approach. Focusing on Seoul’s Jung-gu, Gangnam-gu, Seocho-gu, and Songpa-gu districts, we utilized Street View Images and image semantic segmentation techniques to quantitatively assess the visual elements of street spaces. Volunteers were recruited to rate beauty perception for a subset of SVIs to create a dataset. This dataset was used to train a random forest model to predict beauty perception scores for all SVIs in the study area. Additionally, we examined the impact of streetscape visual elements on beauty perception. The results reveal an uneven spatial distribution of streetscape visual elements and distinct differences in beauty perception between residents and tourists. Korean residents generally perceive greater beauty and show a preference for modernized, bustling commercial environments, while Chinese tourists favor areas that combine greenery with historical and cultural features. Elastic Net Regression results indicate the strongest positive correlation with beauty perception for both groups is with trees, while buildings and walls exhibit negative correlations. These findings provide valuable references for street planning and design, offering scientific support for future urban street evaluation and spatial optimization and promoting the growth of urban tourism.

키워드