How to perceive tourism destination image? A visual content analysis based on inbound tourists’ photos

Abstract

Photos are an important tool for understanding the minds of travelers and characterizing the tourism destination image, and they play a unique role in the construction of tourism images. The wide application of deep learning techniques has brought new opportunities for the visual content analysis of images. This paper proposes a framework for comprehensively analyzing the image of tourist destinations and for targeted tourism planning. Firstly, this paper utilizes deep learning techniques to perceive images from three perspectives: image scene, visual aesthetics and emotional experience, based on photos of inbound tourists in Beijing on the Flickr website. Secondly, spatial visualization analysis of the perceived image is carried out with the help of tourism resource distribution to provide suggestions for tourism planning. The results show that: (1) Beijing's inbound tourism cognitive images can be categorized as food, culture, people, architecture, recreation, natural scenery, city life, animals and infrastructure. (2) The categories of architecture, culture and natural scenery have a higher quality of visual aesthetics and more positive emotional arousal. (3) Spatially, Beijing's inbound tourism image shows the structural characteristics of multi-core aggregation. The findings provide relevant theoretical and practical guidance for destination planning and management.