一种顾及驱动因子重要性的双约束RF-Patch-CA城市扩张模拟方法

城市扩张模拟为实现土地资源合理分配与制定城市发展规划政策提供依据。该文针对传统元胞自动机(CA)在城市扩张模拟中存在城市元胞密集区域团簇现象,耦合随机森林(RF)与基于斑块(Patch)扩张的CA模型,在顾及驱动因子重要性基础上构建基于斑块最大面积和城市扩张总量的双约束RF-Patch-CA模型,并利用该模型模拟重庆主城都市区2010-2017年城市扩张。结果显示:该模型总体精度达97.62%,相比传统的RF-CA、ANN-CA和Logistic-CA模型,Kappa系数分别提高了0.0222、0.0231和0.0245,FoM分别提高了0.0376、0.0391和0.0414;在景观相似度上,该模型相比以上3种模型分别提高了40.92%、41.16%和32.33%,最接近真实情况,而且避免了模拟结果产生城市元胞团簇现象,有效提高了城市扩张模拟精度。

An Urban Expansion Simulation Method of Dual Constrained RF-Patch-CA Considering the Importance of Driving Factors

The simulation of urban expansion provides the basis for the rational allocation of land resources and the formulation of urban development planning policies.When traditional cellular automata(CA)is used to simulate urban expansion,there is cluster phenomenon of dense urban cells.In view of the above shortcoming,a dual constrained RF-Patch-CA model based on the maximum area of patch and total amount of urban expansion is designed by considering the importance of driving factors and coupling the random forest(RF)with the patch-based CA model.The urban expansion simulation results of Chongqing metropolitan area from 2010 to 2017 show that the overall accuracy of the proposed model reaches 97.62%.Compared with the traditional RF-CA,ANN-CA and Logistic-CA models,Kappa coefficient is increased by 0.0222,0.0231 and 0.0245,and FoM is increased by 0.0376,0.0391 and 0.0414,respectively.In terms of landscape similarity,the proposed model is increased by 40.92%,41.16%and 32.33%respectively compared with the above three models,which is the closest to the actual result.This study indicates that the method to constraint the maximum area of patch by considering the importance of driving factors can avoid the phenomenon of urban cell clusters in the simulation results,and effectively improve the accuracy of urban expansion simulation.