土壤遥感分类识别专家系统的结构设计  

Structural design of classification and recognition expert system for soil remote sensing

作　　者：付炜[1,2] 

机构地区：[1]燕山大学信息学院 [2]新疆大学电子科学与信息工程学院,新疆乌鲁木齐830046

出　　处：《生态环境》2004年第1期63-68,73,共7页Ecology and Environmnet

基　　金：国家自然科学基金项目(69662001);乌鲁木齐河流域土地资源动态监测信息系统研究(69662001)

摘　　要：介绍了土壤遥感分类识别专家系统的设计原理与实现方法,阐述了土壤遥感分类识别专家系统的土壤分类决策判断树的构造原理,讨论了土壤遥感分类识别专家知识表示的基本原则和系统知识库中专家知识的组织方式,以及专家系统推理机的设计原理和推理规则的构建方法。系统采用压缩编码方式存储地学专题图形和遥感图像数据,具有数据与图像的存储更新、查询检索、分析处理、特征提取和信息输出等功能。该系统可以对遥感图像进行土壤类型的分类识别,并对分类精度进行监测与评价。用该系统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,并对试验结果进行了讨论与评价。This paper presents design principle and realizable approach of Classification and Recognition Expert System for Soil Remote Sensing (CRES). Meanwhile, the author expounds the building principle of soil classification and decision judge tree for CRES, as well as discusses the basic rules of expert knowledge expression for classification and recognition of soil remote sensing and the organization mode of expert knowledge in knowledge base of the system. The author also elaborates the design principle of inference engine and building method of reasoning rules for expert system. On the basic of mentioned above, the system stores geo-thematic graphics and image data of remote sensing using compressed coding models. Moreover, the system can be used to store, update, inquire, search, analyze, press data image, to extract the feature information from image, and to output new information. The system can be also used to class and recognize soil pattern from remote sensing image, and to examine and evaluate the classification precision. Furthermore, the author makes a test research with the system on classification and recognition soil pattern in the testedregion of Fukang County, situated on the northern foot of the Tianshan mountains, Xinjiang, China. Also, the author discusses and appraises the experimental results.湿地生态系统甲烷排放研究进展

关 键 词：土壤 遥感技术 压缩编码方式 遥感图像 CRES 土壤遥感分类识别专家系统 属性数据库 

分 类 号：S15[农业科学—土壤学] TP79[农业科学—农业基础科学]