Study On The Spatial Changes Of Streets And Alleys In Historical And Cultural Districts Based On Street View Map Recognition

With the continuous advancement of urban renewal,urban physical examination and refined urban management,how to effectively realize the dynamic monitoring and change identification of the spatial quality of streets and lanes in historical blocks has become an increasingly concerned research hotspot.Relying on the development of Internet of Things,big data,unmanned aerial vehicles and other technologies,street view map and street view picture recognition provide new ideas and methods for the research of historical block protection and development.This paper starts from the application of street view map,selects suitable observation points,and builds a historical block research database by obtaining Baidu street view pictures over the years;Taking street view pictures and their analysis and processing as the goal,this paper constructs a street spatial change recognition and evaluation system,analyzes and compares the evaluation results of different evaluation index weight determination methods(questionnaire analysis,AHP analysis)and image recognition methods(ps histogram analysis),and then summarizes the evaluation methods suitable for street spatial change recognition in historical and cultural blocks;On this basis,the Dashilan historical block in Beijing is selected as a research case,and 30 streets and lanes are selected as samples.By identifying and analyzing the street scene picture data,the spatial change law of streets and lanes is summarized.Taking Yanshou Street and Hanjia Hutong as examples,the design strategies and suggestions for street and lane renewal in historical and cultural blocks are put forward.First,determine the block coordinate system,set an observation point every100 m,continuously collect street view pictures,establish and improve the street view picture database of Dashilan historical and cultural block,and collect a total of 438 effective street view pictures.Secondly,ps histogram analysis method and image segmentation method are used to identify the objective elements of each picture.By comparing the two methods,the accuracy and scope of application of different methods are summarized.Thirdly,the street map recognition technology and evaluation system are constructed,and the first-level and second-level indexes are determined.The weight of each index element is determined by questionnaire analysis and AHP analysis,and the evaluation results are calculated and the related influencing factors are analyzed by combining the setting of evaluation scale and the calculation method of spatial change.Finally,Yanshou Street,a residential street,and Hanjia Hutong,a mixed street,are selected as typical streets,and some new design strategies and suggestions are put forward.This paper discusses the ideas and methods based on street view picture recognition and evaluation,which is beneficial to realize the dynamic monitoring of urban block spatial changes.It is found that this evaluation method still has some shortcomings,but it can provide a preliminary basis for large-scale quantitative analysis and evaluation for physical examination of streets and lanes in the old city,renovation of old urban areas and protection of historical blocks.