I~2S101系统是从美国进口的较大型处理图象计算机。本文从理论上和试验上对该机的三个监督分类软件系统进行分析比较,指出它们的优、缺点并根据“平行六面体分类器”和“最小距离分类器”的缺点,提出了对这两个分类器的改进意见,并且引入“多次判决”的概念,进而设计出两个新型的分类器,即“P次判决分类器”和“欧氏线性分类器”。 The I ~ 2S101 system is a larger processing image computer imported from the United States. This paper analyzes and compares the three supervised classification software systems of the aircraft theoretically and experimentally, and points out their advantages and disadvantages. According to the shortcomings of “parallelepiped classifier” and “minimum distance classifier” This paper introduces the concept of “multiple decisions” and then designs two new classifiers, namely, “P-order decision classifier” and “Euclidean linear classifier.”