应用牧区植被的光谱信息特征来识别和描述土壤图中与水分系列有关的制图单元。用单联接聚类分析法把由机载光谱仪获得的光谱数据进行了相对客观和自然的分组。类似的数值分类方法,以往常用于选择某些理化性质以进行土壤分类。本文将1.2-2.4μm范围内的128个光谱段信息作为聚类分析的变量。从显示在枝状图上的各联线,可以看出9个土壤图单元和2个水体间在形态上的相似点或不相似处。从图上还能看到石灰性土,非石灰性土,钠质土和沼泽土等各类土壤以及无堤水体均明显地相聚成类。 The characteristics of spectral information of pastoral vegetation are used to identify and describe the mapping units related to the water series in the soil map. Using single-linked clustering analysis, the spectral data obtained by airborne spectrometer were relatively objective and natural grouping. Similar numerical classification methods, used to select certain physical and chemical properties for soil classification. In this paper, the information of 128 spectral bands in the range of 1.2-2.4μm is used as the variable of cluster analysis. From the lines displayed on the dendrogram, it can be seen that there are similarities or dissimilarities in morphology between 9 soil map units and 2 water bodies. Can also see from the map of calcareous soil, non-calcareous soil, sodium soil and bog soil and other types of soil and no dike body are clearly together into a class.