作者利用美国陆地卫星计算机兼容磁带(CCT)记录的多光谱数字信息,在通用电子计算机上,先后对不同地区用等混合距离分类法(ISOMIX法)进行了自动分类研究。本文对这种方法的分类过程,特别是在研究过程中发现的问题,作了简要叙述和探讨;并作了适当改进,以提高分类的正确性。 Using multispectral digital information recorded by the United States Landsat Computer Compatible Cartridge (CCT), the authors conducted automatic classification studies using isomix distance classification (ISOMIX) in different areas on a general-purpose computer. This article makes a brief description and discussion of the classification process of this method, especially the problems found in the research process, and makes appropriate improvements to improve the classification accuracy.