The Theory And Method Of Human-scale Street Space Planning Based On Street View Image

China's urbanization has shifted from high-speed growth to medium-and high-speed growth,and entered a new stage of development with equal emphasis on scale expansion and quality improvement.Although great achievements have been made in the construction of urban roads in the past few decades,it also brings pressure and challenges to the vitality of urban blocks,the heritage of history and humanity,and the safe travel of cities.Urban roads mainly pay attention to the systematic traffic function,and pay less attention to the slow traffic and street functions and activities mainly in the service blocks.As an important part of urban space,street space,which is composed of many elements such as urban roads,ancillary facilities and buildings along the line,urgently needs to meet more "humanized" needs.Therefore,it is necessary to transform and innovate the existing construction management mode,so as to realize the "humanized" transformation from road to street.The design and management of street space is not only the important content of urban management and control,but also the main entry point of strengthening urban management and control.Strengthening the street construction is an important way to meet the people's demand for public goods and services under the new situation.It can further improve the supply of urban services,stimulate urban vitality,enhance urban cultural connotation and shape urban spirit.Under the background of the transformation of urban planning from "big demolition and construction" to refined planning and management in the past two or three decades,quantitative urban research has attracted more and more attention.The advancement and popularization of information and communication technology in recent years have provided a large amount of new data for quantitative urban research.In the new urbanization period,the rich information about social space and crowd behavior contained in the new data environment is considered as a powerful tool to promote the scientific urban planning and efficient urban governance,which brings new development opportunities for the corresponding academic research and planning practice.At present,it is urgent to establish and improve the policies,regulations and planning theory system of the utilization of stock space resources,change the thinking mode of stock space resources,improve the planning method of the utilization of stock space resources,and innovate the planning technology of stock space.Based on humanism planning perspective,this study to the street space as the research object,build the street-level images as the main data of street space database,using the depth study,GIS and image processing technology and literature research,comparative analysis,practical methods of investigation and the quantitative analysis of the urban space,to build a humanistic yardstick based on street view images street space planning theory,puts forward the research framework of street space planning and process system.Then we take Beijing as an example to study the optimization of street space,which is of great theoretical and practical significance to urban street space planning.Specifically,the main work and contributions of this paper are summarized as follows:(1)On the basis of combing the existing humanism street space planning theory,space scale related theory and the application of street view image in street space,this paper discusses the meaning,theory and method of street space planning on the basis of street view image,aiming at satisfying the different functions and psychological needs of street space.Emphasis of this paper is placed on the supporting role of street view image in the human-oriented street space planning.In this paper,the objectives,principles and key points of the human-oriented street space planning based on street view image are clarified,and the research framework and process system for optimizing street space are constructed.(2)Propose the key technologies and solutions to the difficulties in the human based street space planning method based on street view image are proposed.The key technology of street view image processing mainly includes the extraction of street features and the projection processing of street view image using deep learning technology.On the basis of crawling street view images,based on the research framework and street space planning requirements,using object detection recognition and semantic segmentation technology in deep learning to extract the relevant elements of street view image,and using image processing technology to project the image into fisheye lens or ordinary lens for relevant index calculation and analysis.In addition,it is necessary to quantify and classify other data in the street spatial database,mainly including topological simplification of road vector data,POI data classification and road type judgment.(3)Based on the theory and method proposed in this study,an empirical study is conducted in the central area of Beijing.Aiming at the problems of "more","less" and "chaos" in Beijing's street space,we respectively selected traffic guardrails,street green space and street space quality to conduct empirical research.By analyzing the spatial distribution of relevant evaluation indicators,we can understand the overall status of urban street space,summarize the overall characteristics of street space,combine urban development characteristics and related policy interpretations,explore the mechanism reasons that affect the status quo of street space,and then propose strategies to manage street space.Promote the scientific and efficient management of street space planning.