Exploring City Image Perception in Social Media Big Data through Deep Learning: A Case Study of Zhongshan City

Abstract

Based on Kevin Lynch’s cognitive method of urban image and Weibo’s review data, this study constructs a research framework with three modules as the core: city image structure, city image types, and cultural service evaluation. First, the geospatial information carried by comments is analyzed by GIS to obtain the image structure of the city; second, the picture data in the comments are divided into image types and the type ratio is calculated by the image semantic segmentation method based on deep full convolution neural network. Finally, the text data in the comments are extracted from the semantic word frequency analysis to evaluate the cultural service perception index words of the city image and combined with the analysis of the city image structure and the city image type so as to obtain the integrated comprehensive perception of the city image. The research shows that the introduction of big data and deep learning methods into city image research can make up for the shortcomings of traditional research samples, expand the dimension and breadth of urban cognition, reveal the social, cultural, and functional characteristics of the city, and is an important supplement to the five-element model of city image depicting the material form of the city. In addition, the results of the empirical study, taking Zhongshan City as an example, have implications for the realistic urban spatial planning, urban landscape design, and tourism industry layout of Zhongshan.