Cellular Automata Fine Simulation Of Urban Land Use Change Using Remote Sensing Images And POI

Land use/cover change reflects the interaction between human activities and the natural environment,and has always been one of the core issues in the field of sustainable development.With the rapid development of urbanization,the problems of unreasonable development of land resources and deterioration of ecological environment emerge one after another.The contradiction between human and land is becoming more and more intense.Cellular automata(CA)has the ability to simulate the spatiotemporal dynamic changes of complex systems,thus becoming one of the powerful tools to simulate urban land use changes.Therefore,carrying out urban land use change simulation using cellular automata can reveal the complicated laws of urban land use evolution.This will be of great significance for optimizing the spatial structure of urban land use and realizing sustainable urban development,both theoretically and practically.Remote sensing images mainly extract natural attribute information such as farmland,forest and water,and have become an important data source for urban land use change research.However,related studies show that the method of extracting land use information using remote sensing image has the problem of too macro scale of classification results.That is difficult to reflect the complex social functions in the land,such as residential function Can,commercial function,etc.With the rapid development of the network society,multi-source geographic big data has been widely used in the research of urban land use function identification.It has the advantages of low acquisition cost,large amount of data,rich variety and fast update speed.In addition,in the process of fine simulation of urban land use change using cellular automata,the scale selection differences of model input factors such as cell size,neighborhood size and neighborhood type will have an uncertain impact on the simulation accuracy.Therefore,how to reasonably apply multi-source geographic big data to urban land use change simulation and optimize the parameters of cellular automata model is an urgent scientific problem to be solved in the current fine simulation of urban land use change.This will improve the degree of refinement of cellular automata simulation,and simulate the results reflecting the complex characteristics of urban land use micro pattern evolutionBased on the above research background,this paper focuses on the scientific problem of fine simulation of urban land use.After sorting out relevant research results at home and abroad,we study from three aspects:fine identification of land use information,selection of optimal scale combination and fine simulation using cellular automata(1)Urban land use fine identification based on remote sensing image and POI.Based on the remote sensing image as the basic data source,the macro classification pattern of urban land use is extracted by supervised classification method.On this basis,POI geographic big data and the kernel density analysis method is used to identify the urban land use function in the block unit.Then we obtain the micro classification pattern of urban land use,which provides the data basis for the fine simulation of urban land use change.(2)Quantitative evaluation of scale uncertainty of cellular automata for fine simulation process.In order to measure the model uncertainty reasonably,this paper comprehensively considers the factors that affect the simulation accuracy,such as cell size,neighborhood size and neighborhood type.It constructs a coupling evaluation index system integrating the overall accuracy,kappa coefficient and FOM index.The response surface method is used to quantitatively evaluate the scale uncertainty of cellular automata model in the process of fine simulation.Then the optimal scale combination is scientifically selected to improve the reliability of fine simulation results of urban land use change.(3)Urban land use fine simulation using cellul.ar automata model.Based on the refined classification data and optimized model parameters,the driving factors affecting urban land use change are selected to make suitability atlas.The simulation model of urban land use change based on cellular automata model is constructed to carry out the fine simulation of urban land use change and reveal the complex law of urban land use micro pattern evolution.Taking the rapid development area of Wuhan City as an example,the application research and model validation were carried out.The results show that the overall accuracy of confusion matrix between the fine classification pattern of urban land use and the actual classification status map in 2020 is 76.33%,and the kappa coefficient is 0.7396.It can be seen that the kernel density analysis method used in this paper can be used for urban land use function identification,and the identification results are more refined.The relationship between influencing factors and simulation accuracy is analyzed by response surface method.The optimal scale combination is obtained as follows:cell size is 60m,neighborhood size is 5×5,and neighborhood type is von Neumann.In 2020,the precision of urban land use change fine simulation results and classification results were verified.The standard kappa,location kappa,partition kappa and random kappa were 0.7973,0.8625,0.8625 and 0.8188,respectively,and the overall precision and FOM index were 0.8331 and 0.3207,respectively.This indicated that it was feasible to carry out the fine simulation of urban land use change in Wuhan rapid development area by using CA Markov model and the simulation accuracy is high.The simulation results in 2025 show that the land use pattern in the rapid development area of Wuhan will show a trend of decreasing fragmentation,decreasing spatial heterogeneity and increasing shape consistency from the center to the periphery.Using remote sensing image and POI data,this paper puts forvard a set of technology and method system of cellular automata fine simulation of urban land use change under the cooperation of remote sensing image and POI.The method system can effectively control the accuracy and reliability of cellular automata simulation results from two aspects of data source and model parameters,and effectively improve the degree of refinement of urban land use change simulation.The results can provide some scientific reference for the compilation of territorial and spatial planning and sustainable urban development.