应用光谱资料对土壤和土地利用进行数值分类

土壤的波谱反射特性是土壤遥感数据信息处理和计算机自动识别分类的物理学基础.我们曾以云南腾冲光谱试验区的室内外测试资料,应用聚类分析进行了土壤和土地利用分类的尝试.本文拟用同样资料,应用主组元分析进行分类识别,以期为土壤遥感的数据处理,提供一个较为有利的途径.主组元分析已在包括土壤和遥感在内的各个科学技术领域中得到广泛应用.关于主组元分析的原理和计算步骤,本文不另赘述.

NUMERICAL CLASSIFICATION OF SOILS AND LAND USE BY SPECTRAL DATA

Based on the spectral data obtained from laboratory and field experiments in Tengthong spectral test area,the automatic identification and classification of soil types and hand use patterns had been carried out by the use of the principal component analysis.Two principal component equations based on the laboratory data can be used to identify soil types developed from different parent materials and of different land use patterns:Another two principal component equations based on field data can be used to distinguish different land use patterns such as fallow,bare land,wheat field and water etc.