This paper mainly explains two methods about how to improve the accuracy of remote sensing classification. The first
method is composite hierarchical classification with multiple information sources, assisted by GIS and based on
statistical interpretation. The other one is analysis on background parameters. The experiment shows that they not only
improve the methods of the conventional remote sensing classification, but also raise its accuracy.
The article reports on the research work about applying remote sensing technology to conduct agricultural natural
resources investigation and dynamic monitoring, and soil conservation information system management, and introduces
major achievements at various stages. After 35 years of research in Shanxi Province agricultural natural resource
investigation and dynamic monitoring have gone through the initial experimental stages of macro, micro and the
combination of fixed location with its property and quantity, and have now developed into mature quantization research
stages of fusion system of combing remote sensing, Geographical Information System and Geographical Information
Science, so as to provide practical research technology for agricultural natural and environment management and
decision making. With rapid development of science and technology, deepening of the remote sensing research, we will
gradually set up Numeric Agriculture in Shanxi Province and decision making and information management system of
resources and environment monitoring, improvement and exploration on county, prefecture and provincial scales, so as
to put Shanxi agricultural natural resources and environment management on the scientific way. It has a great
significance in a long term.