【摘要】 在 1 98 8～ 1 996年 5个时段景观遥感制图的基础上 ,利用细胞自组织模型对深圳市龙华地区城镇建成区的动态变化过程进行了模拟研究。模型共选择了高程、坡度、距最近干线公路的垂直距离、斑块相邻度指数和形状指数等 5个变量参与非城镇用地单元的综合转移概率计算 ,利用 1 988～ 1 990年的城镇扩张结果确定不同变量的权重参数 ,利用 1 992、1 994和 1 996年的数据进行模型验证。结果表明 ,细胞自组织模型可以很好地反映模拟起始年份已有城镇斑块的增长过程 ,但是模型不能自动判定新出现的城镇斑块并模拟其动态变化过程 ,成为该模型应用的一个重要限制性因素更多还原

【Abstract】 Based on Multi temporal remotely sensed landscape mapping data,cellular automata model,belonging to the individual based landscape model,was used to simulate the changing process of town landuse for Longhua area,Shenzhen City.The cellular model is constructed to calculate the integrated probability of a particular non town cell,which is just near a town cell,being assimilated into town component.As for non town cell,its transforming probability is determined by the composite effect of five influence factors,which are elevation,slope,distance to the nearest highway,the neighborhood index to the town area component,and the shape index of nearest town patch within a special searching range. Simulation results indicated that the rapidly changing of town area would continue till the end of 2005.Because of the amalgamation process,a lot of town patches merge into big ones from 1990 to 2005 and made the patch amount decreasing significantly,and the change of patch shape is just contrary.During the simulation period,town landuse kept macroscopic heterogeneity characteristics.There still remains some deficiency in the cellular automata model,because it can not tell out intelligently when and where a new town area cell will arise.更多还原