Font Size: a A A

Spatio-Temporal Dynamic Simulation Of Urban Spatial Three-Dimensional Growth Based On XGBoost-CA ——A Case Study Of Dalian,China

Posted on:2022-12-22Degree:MasterType:Thesis
Country:ChinaCandidate:C F GuoFull Text:PDF
GTID:2480306782468514Subject:Computer Software and Application of Computer
Abstract/Summary:PDF Full Text Request
Following the pace of my country's high-speed economic development,the rapid expansion and evolution of large and medium-sized urban areas in China,and the spatial pattern succession of its structure in the horizontal dimension has been unable to match the requirements and restrictions of the current national conditions,people's conditions and other practical conditions.The trend of urbanization is becoming more and more intense,so the research on the three-dimensional growth simulation of urban space and other related fields from this perspective needs to be enriched,and the substantive urban evolution law that is more in line with the actual development trend needs to be developed urgently,and its significance is particularly far-reaching and crucial.The research object of this thesis is the urban building group with the floor height distribution as the core,and a three-dimensional index system is constructed to describe the form replacement in the urban three-dimensional space in detail,and to explore the internal law of the development and transformation of the urban three-dimensional space.The three districts of Zhongshan,Xigang and Shahekou are used as models to simulate the three-dimensional spatial development of the main urban area of Dalian,which has an indispensable reference for the promotion and promotion of the research on urban high-dimensional spatial form.The brief content and conclusions are as follows:(1)Construct a cellular automata urban three-dimensional evolution model based on the extreme gradient boosting tree algorithm(XGBoost).Based on cellular automata,through the simplification and understanding of the characteristics of urban form evolution,based on cellular automata theory,gradient boosting tree classification algorithm,and supplemented by local climate zone theory(Local Climate Zones,LCZ),a cellular automata is constructed.The automata-based urban three-dimensional evolution simulation and prediction model(XGBoost-CA)elaborates on the internal influencing factors of the model,the neighborhood and the transformation rules that limit the state of individual cells.Firstly,according to the obtained two-dimensional spatial distribution of buildings in the main urban area,as well as the building height and density,the XGBoost-CA model is used to calculate the mutual conversion probability between various building plot types under the combined action of different influencing factors;The comparison between the simulation results and the reality,namely the verification of the accuracy of the model,proves that its simulation computing capability can be relatively accurate for the distribution of buildings and the evolution of levels in the main urban areas of coastal cities with complex terrain represented by Dalian in the past ten to twenty years.Prediction research on the three-dimensional form of the city;finally,based on the application of the finished product simulation model,the three-dimensional spatial form succession of the three districts of the main city of Dalian in five years is estimated.By using the XGBoost-CA simulation model,the distribution of buildings in Zhongshan,Xigang and Shahekou in 2020 and the temporal and spatial evolution of comprehensive types of buildings are simulated.Combined types to make predictions.(2)Constructing the urban three-dimensional trend surface based on the height of the building,analyzes and studies the characteristic indicators of the succession of the three-dimensional spatial form in the urban area of Dalian,and summarizes its overall development law.Relying on the calculation and analysis of the regional spatial pattern distribution of urban building groups,the effective performance reflects the spatial pattern distribution characteristics of various types of buildings.And further combined with the introduction of the floor height of the buildings in the study area and the base area,the urban spatial form landscape measurement method was established,which provided the basis for quantitative analysis of the evolution of the urban three-dimensional spatial form.The main achievements and significance are as follows.The evolution process of each parameter index of the three-dimensional spatial form in the three districts of the main city of Dalian in the time dimension was calculated and calculated,and the regular fishing net was created and divided to determine the three-dimensional spatial form of the three main districts of the main city of Dalian.The scale sensitivity of Dalian is tested,and the temporal and spatial differentiation of three-dimensional spatial form succession in the three districts of the main city of Dalian is analyzed.(3)The three districts of Zhongshan,Xigang and Shahekou,the main cities of Dalian,showed a significant growth trend in the vertical and vertical directions in the five years from2010 to 2015.At the same time,the total number of buildings of high-rise and above types also increased.showed a marked increase.The three-dimensional urban spatial form indicators of the three main city districts in the five years from 2015 to 2020 are all scale-sensitive,following the continuous expansion of the research scale,and various areas reflected by the urban three-dimensional spatial form indicators Sexual differences are also increasing.As the differences within the study area become more and more significant,the similarity between the three-dimensional spatial form and landscape of Dalian city in each area gradually decreases along with the progress of time.
Keywords/Search Tags:Urban 3D spatial form, XGBoost-CA, dynamic evolution, simulation model, Dalian City
PDF Full Text Request
Related items