CAJ | 학술논문

提出了一种基于GIS的地理元胞自动机模型框架:SimUrban,用于城市发展和演化的模拟与预测.该框架基于面向对象技术,在GIS环境下利用VS.NET开发而成,可以集成遥感和GIS数据以及新的转换规则和地理CA模型,从而模拟城市演化并进行精度评定.以上海市嘉定区为例,在SimUrban环境下利用基于主成分分析(PCA)的地理CA模型模拟了该区域1989-2006年城市发展和演化过程.
제출료일충기우GIS적지리원포자동궤모형광가:SimUrban,용우성시발전화연화적모의여예측.해광가기우면향대상기술,재GIS배경하이용VS.NET개발이성,가이집성요감화GIS수거이급신적전환규칙화지리CA모형,종이모의성시연화병진행정도평정.이상해시가정구위례,재SimUrban배경하이용기우주성분분석(PCA)적지리CA모형모의료해구역1989-2006년성시발전화연화과정.
In this paper,a novel GIS-based geographic CA modelling framework,SimUrban,is proposed to simulate and project urban expansion and development. Based on object oriented programming technology,the framework is developed by using Visual Studio.NET in GIS environment,which enables the wide usage of RS images and GIS data.The proposed framework composes three parts:1)data acquisition,using SimUrban,ArcGIS,and Envi,the vector topographic maps and RS images are integrated;2)new transition rules of CA,such as partial least squares regression method,least square support vector machines,particle swarm optimization,kernel principal component analysis,and kernel logistic regression method,are allowed to imbedded;3)using this framework,urban expansion and land use/cover changes are simulated with assessment of the results by the error matrix method.The urban development and growth of Jiading District,Shanghai,from 1989 to 2006 was successfully simulated in SimUrban with the CA model based on principal component analysis (PCA).Simulation accuracy of non-urban category is 82.04%,urban category is 70.23%,the overall accuracy is 76.55%,and the Kappa coefficient is 52.58%.The implementation of Jiading District has demonstrated that the modelling framework can generates patterns match well with the actual urban morphology.