City Image Element Recognition Based On POI And OSM Data

Image of the city is the subjective impression of the observer on the image of the city.It is an important part of urban geography and planning studies,and has important guiding significance for shaping the unique empty city space.The study of traditional urban imagery mainly relies on questionnaires and sketching of imagery.It has the problems of high cost,low update frequency and limited data coverage,which cannot meet the quantitative research of cities in the information age and the needs of urban economic development.In recent years,with the advent of the era of big data and the development of Internet technology,more and more quantitative research results have been carried out in cities through emerging data such as POI,open map data(OSM)and social media.Applying to urban imagery research will be a viable approach.In view of this,this paper proposes a method for identifying urban image elements based on POI and OSM data.The main research contents are as follows:(1)The city image element recognition system based on POI and OSM data is studied.On the basis of summarizing the theoretical basis and application practice of urban imagery at home and abroad,this paper deeply analyzes the Volunteer Geographic Information(VGI)and its application in urban research.In view of the shortcomings of traditional data in urban research,it proposes a city based on POI and OSM data.Image element identification system framework.Based on the improved DBSCAN clustering algorithm under the road network constraint,the Denis-Graphbased POI density change analysis and the LDA short-text information classification based on the LDA theme model,the urban image element recognition framework was constructed.(2)A cognitive method of urban image structure is proposed.By clustering the POI data under constraint conditions,cluster clusters are generated to identify the urban image area elements;the core density analysis principle is used to draw the DensiGraph curve of the POI data,and the natural and administrative divisions are superimposed to identify The urban image boundary element;on the basis of thinning the POI data,the city image symbol element is identified,and the processed OSM road network data is combined to identify the city image node and the path element.(3)A cognitive analysis model of urban image element types is constructed.Through the LDA topic model,the topic information of the attribute information contained in the POI data is classified,and the city image element type analysis under different topics is performed.At the same time,the clustering results of different topics are superimposed with the above-mentioned element structure to explore the spatial characteristics of different thematic city images,and summarize the shortcomings.(4)Finally,taking Jinan City as an example,the model verification and analysis prove its important significance and potential value in urban research and perception.The research results show that the research methods used in this paper can be used to identify urban image elements,explore urban image structure,and recognize urban image types,thus promoting the shaping of unique urban images and the healthy and sustainable development of cities.