Font Size: a A A

Research On Location Selection Of Future Urban Parks Based On GIS And Machine Learning

Posted on:2021-02-28Degree:MasterType:Thesis
Country:ChinaCandidate:H T QinFull Text:PDF
GTID:2370330647450993Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
As the world's largest developing country,China has experienced the largest and fastest urbanization process since reform and opening up.In order to meet the production and living needs of the population gathered in the city,it will have a serious impact on the ecological environment.Urban parks are an important part of the city's functions,and they can provide leisure and recreational functions,and play an important role in promoting sustainable economic development and enhancing urban livability.However,the quantity and quality of parks in many cities are currently difficult to meet the needs of residents.With the further expansion of cities in the future,the urban population's demand for parks will further increase,and new parks are urgently needed to alleviate the contradiction between supply and demand.Most of the current research on urban expansion simulation is based on the combination of a single machine learning algorithm and cellular automata,and there is a lack of comparison between multiple models.At the same time,the study of park location only focuses on the current supply and demand relationship within the city,and less consideration is given to the potential demand of the park for future urban development.In order to match the green space with the fast-developing city and protect the residents' demand for parks,four district-level administrative units(Xinbei,Zhonglou,Tianning,and Wujin)in Changzhou City in the Yangtze River Delta region were selected as the research Area.In a long period of time in the future,the urban construction and economic development of Changzhou City will still maintain a relatively fast level,so the research has certain representativeness and reference significance.The paper is based on multi-source spatio-temporal data,using multiple GIS spatial analysis methods,multiple machine learning algorithms,and cellular automata to simulate the future urban land use spatial pattern and construct a city park site selection model,and then give decision suggestions for park site selection planning.The contents and conclusions of research are listed below:(1)Analysis of historical changes and layout of urban parksFirst,based on historical data,the historical changes of urban parks in the study area were analyzed in 1985,2000,and 2015;secondly,spatial analysis methods such as nuclear density analysis,cellular density analysis,and buffer zone analysis were used to analyze the park's spatial layout and service scope.Analysis;Finally,based on the distribution of existing parks,the current intensity of demand is comprehensively calculated through the three factors of park core density,distance from the park,and population density.The results show that the number of parks in the central urban area of the study area has been continuously increasing.The dense areas of the park are mainly distributed in the central urban area,and the density gradually decreases from the center to the outside,showing a more obvious "circular ring" distribution.According to the "15-minute life circle" concept,the park service range of 1km and 3km radius is set.The service coverage rate of the buffer zone with a radius of 1km is 29.07% and the service coverage rate of the buffer zone with a radius of 3km is 79.39%.Overall,the 3km service range of Changzhou Park is more comprehensive and can better meet the needs of residents,but it is still difficult to meet the current needs in regional parks in northern Xinbei District and southern Wujin District.(2)Simulation of future urban land expansion under the constraints of land ecological suitabilityBased on the spatial-temporal evolution of land use in the study area,the characteristics of urban spatial-temporal evolution are analyzed.Combining the three conditions of natural geographic conditions,human activities and ecological constraints,the results of land ecological suitability evaluation in the study area were obtained.Using a variety of machine learning algorithms combined with cellular automata,adding land ecological suitability as a development constraint,comprehensively constructing a future urban land expansion model in Changzhou City,and obtaining key areas for urban land growth in 2030.The results show that The land use types in the study area have undergone tremendous changes in space and quantity,and a lot of agricultural land has been converted to construction land.The ecological suitability of the land in the study area is generally good,and the ecological suitability of the land in the central part of the New North District and the southern part of the Wujin District is high.The total area of urban land in the study area is expected to be about 538.47 square kilometers in 2030,an increase of 147.94 square kilometers compared with 2015.The newly added areas of urban land are mainly distributed in the northwestern part of the Xinbei District,around the central urban area,and around the township centers in the south of Wujin District.Overall,the future development potential of the urban area in the study area is large.(3)Location selection of future urban parksBased on the current demand intensity,land ecological suitability and future demand intensity,a spatially balanced urban park location model is constructed.The specific location of the park construction is determined by the suitability evaluation of the park construction and the greedy algorithm model.Through marginal benefit analysis,suggestions for planning decisions for future urban park construction are given.The results show that the suitable area for park construction accounts for 20.47%.The suitable area is mainly distributed in the northern part of the Xinbei District,around the existing urban land in the central urban area,and the southern part of the Wujin District.Combining the results of the cumulative histogram and the line chart,the number of newly built parks is 7 and 14 is a "catastrophe point" with a large change.It is recommended that the government should choose 7 new parks as the first stage of the planning goal in the construction of future urban parks,and the overall service coverage of the park will reach 85%.During the second phase of the planning,the construction of the next 7 parks,after the cumulative construction of 14 parks,the overall service coverage rate can reach 95%.After the future expansion of the city,economic strength and residents' leisure and recreation needs further increase,then consider the site planning of the new park.In summary,based on the current demand intensity,land ecological suitability evaluation and future demand intensity,the construction of a spatially balanced urban park location selection decision-making model is established to improve the urban park location theory system.At the same time,it provides countermeasures and measures for the future urban park space selection,effectively improves the service coverage of parks in the study area,thereby improving residents' recreational comfort and enhancing the livability of Changzhou City.
Keywords/Search Tags:Park Location Selection, Land Ecological Suitability, Urban Land, Machine Learning, Changzhou City
PDF Full Text Request
Related items