基于模糊相似性之概率区间三维土壤地图
Three-dimensional soil map of probabilistic interval based on fuzzy similarity measures
-
摘要: 一般工程上土壤分类常使用定性方法,而无分类量化讯息以提供空间推估使用,因此常使用模糊k-means(FKM)法,本文尝试将FKM结果与美国统一土壤分类法进行比较,另提出以模糊相似度来改善FKM应用于工程土壤分类,依据概率分配来设定临界值求得土壤分布之置信区间,文中并以一范例地区说明绘制三维土壤地图之过程。Abstract: In civil engineering,the deterministic system is usually used for soil classification,but it can not provide the quantitative information for the spatial interpolation.The fuzzy k-means is often used to quantify the soil classification but no comparison with other systems had been given.In this paper FKM was compared with USCS to validate its abilities of classification which have been improved with a formula of fuzzy similarity measure.Based on the probabilistic distribution,the confident interval of distribution of soil groups could be estimated.The process of establishment of 3D soil map was illustrated in a practical case.