Abstract:In the context of the digital age, the development of artificial intelligence technology has broadened the access to data. In this study, we took 5A-level tourist attractions as the research object, used the directory and attribute data of the official platform and the text data deeply synthesized by ChatGPT to mine tacit knowledge, used descriptive and inferential mathematical statistics methods to identify the key spatial image elements of tourist attractions, and analyzed the relationship between congenital conditions and acquired intervention and spatial imagery. The results show that: (1) Markers are the primary basis for distinguishing different scenic spots in tourists’ impressions, and they are also the key to obtaining higher tourist image evaluation for individual scenic spots. (2) Landmarks, areas, and roads constitute the basis and main body of the overall spatial image of 5A-level scenic spots. (3) The influence of acquired design on imagery is significant, and the creation of markers and regional divisions are the key points in the acquired design process. This study can improve scenic spot planning, design, and high-quality development and promote refined image publicity and management.