珠江口土地利用变化推测的CBR方法  

The CBR Approach with the Inference of Land Use Change at Pearl River Estuary

作　　者：温伟[1,2] 杜云艳 王春晓[1,2] 

机构地区：[1]中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室,北京100101 [2]山东科技大学测绘科学与工程学院,山东青岛266510

出　　处：《山东科技大学学报（自然科学版）》2009年第3期91-98,共8页Journal of Shandong University of Science and Technology(Natural Science)

基　　金：国家“863”计划探索导向课题(2007AA12Z222);资源与环境信息系统国家重点实验室自主创新团队计划项目(088RA400SA);中科院知识创新项目(kzcx2-yw-304)

摘　　要：传统CBR方法在应用到地学领域时存在着明显的缺陷,急需在CBR的表达模型及推理算法上进行拓展。而土地利用变化影响因素众多,且具有非常复杂的空间或区域特征,用单一的统计方法或动态模型难以进行完全的定量分析。基于此,提出了土地利用变化的CBR方法研究。首先,在概述土地利用变化当前定量研究方法基础上,提出了利用CBR方法进行土地利用变化的分析与推理的思路;其次,针对土地利用变化的空间特性及所隐含的空间关系特性,给出了进行土地利用变化推测的CBR具体方法,包括CBR方法进行土地利用变化案例的表达,土地利用变化案例间内蕴的空间关系的抽取,以及考虑了案例间空间关系的CBR相似性推理方法等;最后,以珠江口西岸土地利用变化比较典型的珠海部分区域进行了CBR方法的应用试验研究。The obvious defects exist in application of the traditional method of CBR in geoscientific field, and it is urgently needed to expand CBR expression model and the reasoning algorithm. And, there are many factors affecting the land use change and they are of very complex spatial or regional characteristics. It is difficult to complete quantitative analysis by using single statistical method or dynamic model. Therefore, this paper proposed a CBR method for studying the land use change. Firstly, in outlining the current land use change on the basis of quantitative research methods,the paper brings forward the CBR method for analyzing and reasoning the land use change; Secondly, according to the spatial characteristics and the spatial relationships implied in the land use change, the concrete method for using CBR in inference of land use change is given out. The method includes as follows： the method of CBR expressing the cases of land use change, the abstraction of intrinsic relationship between regions where land use changed, as well as the method of reasoning the similarity of spatial relationship between cases. Finally, we did the CBR test of the typical Zhuhai land use change in an area of west bank, a part of the Pearl River estuary.

关 键 词：人工智能 案例推理(CBR) 土地利用 空间关系 

分 类 号：F301.2[经济管理—产业经济] TP181[自动化与计算机技术—控制理论与控制工程]