基于GIS和单元簇的模糊逻辑推理及其在区域矿产预测中的应用

论述了 GIS支持下的一种基于“单元簇”概念和模糊逻辑推理的多元地学信息综合分析方法及其在区域矿产预测中的应用。针对以往矿产定量预测中的单元划分方法对空间信息利用不足的问题 ,用单元的空间组合 (“单元簇”)代替单元作为定量类比的基本单位 ,从而能较充分地利用地质变量的局部空间结构信息 ;将单元作为 GIS区图元 ,利用 GIS空间分析功能实现对单元及单元簇的管理和操作 ;建立两个层次即变量对单元和组成单元对单元簇的模糊推理规则 ,经两次模糊推理计算出所有未知单元的找矿有利度 ,为进一步圈定找矿远景区提供了基础 ,并以新疆康古尔塔格地区金矿预测实例说明了其应用效果

GIS-BASED CELL CLUSTERFUZZY LOGIC DEDUCTION AND ITS APPLICATION IN REGIONAL MINERAL RESOURCES PREDICTION

This paper presents a fuzzy logic deduction method forquantitative prediction of regional ore resources based on the concept of “cell cluster” and supported by GIS. With cell clusters (groups of cells spatially associated) we can ensure more sufficient use of the spatial structural information of geo-variables. Cells are taken as GIS polygon elements so they as well as the clusters can be manipulated through the spatial analytical functions of GIS.Two levels of fuzzy reasoning rules are built. The first level rules determine in what extent a geo-variable makes a cell “good”(i.e., qualified for a specific member of the model cluster); The second level rules determine in what extent a cell makes a cluster “good”(i.e.,favorable for ore prospec- ting). Through the two-leveled fuzzy deduction the favorability of any unknown cells can be produced and then the favorable prospecting areas can be delimited. A case study of gold mineral resources prediction in Kangurdage region, Xinjiang, is present.