The Image Characteristics Interpretation and Influencing Mechanism of Lingnan Towns from the Perspective of Internet Media

ZHAO Miaoxi; XU Min; YAO Sheng; XIE Tengxiao; LIU Quanyu; YIN Wenbo

doi:10.3969/j.issn.1000-0232.2025.11.009

Abstract：The development of social media and artificial intelligence in the Internet era has promoted quantitative research on urban image.To reveal the image characteristics of Lingnan towns and their influencing mechanisms from the perspective of the mobile Internet,this study focuses on 136 towns in Guangdong,Guangxi,and Hainan provinces with a large number of videos on the Douyin platform.Integrating multimodal data,such as keyframe image tags for Douyin videos and Bilibili video comment texts,this study systematically reveals the characteristics,type differences,and action mechanisms of the urban image of Lingnan towns in Internet media through image recognition,text analysis,structural equation model,and other methods.The research finds that the selective presentation of fragmented scenes in the virtual space constructs the main image expression of typical Lingnan towns.Most high-frequency image points are outside the town area,indicating that the built-up areas in Lingnan town are insignificant.The secular life scenes depicted by the communication subjects are superimposed on natural landscapes,such as characteristic mountains and leisure islands,which become the core components of the network imagery of Lingnan towns in the real environment.From the town's perspective,the network imagery is mainly influenced by population concentration,urban vitality,and transportation accessibility.Among them,population concentration and urban vitality have significant positive effects,while transportation accessibility can facilitate the network imagery positively.This study provides some empirical support for understanding Lingnan town images in the Internet media environment and explores ways to expand urban image theory.