介绍了干旱区土壤遥感分类识别推理决策器的设计原理与实现方法。在用TM遥感图像对土壤类型进行非监督分类的基础上，建立了正向推理与逆向推理相结合的推理机制，对土壤类型进行分类识别决策。用知识表示的产生式规则与框架式规则相结合的数据结构表示土壤学专家的土壤分类识别知识。用像结构模式建立了土壤分类识别的规则，构造了土壤分类判决树，并用典型像例模式进行了各类型土壤判据文件的组织。用该方法对新疆天山北麓阜康试验区的土壤分类识别进行了试验研究。结果表明，该方法分类精度可靠，为干旱区土壤分类识别开辟了一条新的途径。

This paper presents design principle and realizable approach of Classification and Recognition Inference Decider for Soil Remote Sensing (CRID) in arid land. On the basis of non_supervising classification for soils with TM images, the author discusses the reasoning mechanism by combining the direct inference combined with inverse reasoning for soil classification and recognition decision. The author also expresses soil classification and recognition knowledge of expert in soil science using data structure of producing rule linked with frame rule for knowledge expression in the CRID. Furthermore, the author make up the rules of soil classification and recognition with image structural model, and builds decision tree of soil classification in the CRID, and organizes files of decision for soil classifications with typical image case model. With these methods, the author carries out a test research on classification and distinguishing for soil in test region of Fukang Counry, situated on the northern foot of the Tianshan Mountains, Xinjiang Province. And the test result shows that the approach mentioned above has a high reliable precision, and it reclaims a new way for classification and recognition of soil in arid land.