基于社交媒体数据的徒步道景观要素识别与偏好模式OA

[Objective]This study aims to identify landscape elements of hiking trails based on social media data,and reveal the characteristics and patterns of tourists'visual preference.[Method]The Shili Langdang Trail in Hangzhou is selected as the case.Tourist-generated texts and geo-tagged images are collected,and a multi-source analytical framework integrating qualitative analysis,sentiment analysis,and image semantic segmentation is constructed.[Result]Landscape elements perceived by hikers are classified into eight categories:natural landscapes,urban landscapes,service facilities,ancient trails,rural landscapes,landscape architecture,historical relics,and hikers.Natural landscapes are dominant,significantly correlated with positive sentiment.Four visual preference patterns are further identified to include nature-oriented aesthetic type,place-narrative and nature-synergy type,socio-symbolic human-landscape integration type,and route-characteristic immersive experience type.[Conclusion]Social media data can characterize the visual focus and emotional tendencies of hikers,and reveal the linkages between landscape element group and experiential perception.The results can provide a basis for landscape element allocation and recreational space optimization of hiking trails.Future research can integrate field investigation and subjective evaluations to further validate and improve the evaluation framework.Technologies such as virtual reality and eye-tracking can be incorporated to explore the micro-mechanisms of visual attention and emotional responses,and to extend towards multisensory experience,thereby supporting refined design of linear landscape spaces.