Research On The Characteristics Of Internet-famous Cities Based On Image Perception

The process of global integration has intensified competition between cities.A well percieved city image and appropriate publicity methods are of great importance to improve the popularity and attractiveness of cities.The advent of the Internet era has made urban marketing more diversified.New media forms such as short videos have explored and displayed urban characteristics through languages easy to understand,turning the city's image from cold to friendly,and at the same time giving birth to a number of “Internet-famous cities”.Thus,the study of urban image is even more urgent in modern times.Understanding people's core concerns about a city,and grasping the true perception of the city among its stakeholders,will help identify the strengths and weaknesses of city construction,and take targeted measures to consolidate the city's image.With the development of big data and corresponding computer technology,the materials and methods of image research have been greatly expanded,and the cognition of image connotation and its attributes has been continuously enriched,resulting in many inspiring research perspectives.In view of this,this paper proposes a method for identifying and analyzing urban image elements based on user generated big data using semantic mining.The main contents include the following aspects:(1)An image element recognition model based on natural language processing is constructed.In connection with the current research situation of small sample,low efficiency,simplicity and homogeneity in image text analysis,this paper proposes an image perception model suitable for the analysis of big text data.Text semantic mining methods including keyword extraction based on tf-idf algorithm,word vector model based on word2 vec,keyword clustering based on vector cosine distance,and image class coding based on grounded theory are adopted to established an image element recognition model.(2)Expanded the cognitive dimension of city image.On the basis of studying the similarities and differences in contents,methods,data,perspectives and trends of urban space image research and destination image reserach,as well as the evolution process of image connotation,the structure of image is obtained according to image recognition model,which includes three image dimensions: entity,space and psychological perception.And emotion analysis is introduced to realize the comprehensive perception of cognitive image and its corresponding emotion.(3)Take Chongqing,Chengdu and Xi'an as examples to conduct empirical research for the purpose of verifying the feasibility of the model.The overall characteristics of image perception in three cities are analyzed from four perspectives: kinds of image,perception intensity,emotional perception,and spatial distribution.And further explore of distinguishing image,advantages and disadvantages in image perception are conducted through image type comparison between cities,IPA analysis,and social network analysis based on image co-occurrence.The results show that the method proposed in this paper is suitable for the excavation of urban image structure and image characteristics,and can provide reference for the improvement of urban image and long-term development of the city.