Remote Sensing Image Change Detection Method For Geographic National Condition Monitoring

For the applications of geographic national conditions monitoring important engineering, this article focuses on researching high-resolution remote sensing image change detection methods and technology, establishing the change detection process and change detection software system, and then realizing the initial application of the method and system, providing technology and software platform for geographical elements change detection, including the following aspects:(1) Study on the change detection method based on the multi-level decomposition model. This article analyses the characteristics of the typical objects, and then establishes a multi-level decomposition, uses decomposition unified template model on the before and after remote sensing images. With the minimum decomposition sub-block as a basic processing unit to detect, explore the decomposition idea of the change detection which based on multi-level decomposition.(2) The changed information’s prediction based on fuzzy rules. Set the determined threshold, and the entire image is classified as three regions which are determined change, determined no change and fuzzy change. The fuzzy change region is classified in accordance with the fuzzy judgment rules, auxiliary manual interpretation, thus obtain the change regional information.(3) Object-oriented change information extraction. By the means of adaptive threshold analysis, change detection result obtained by the fuzzy judgment is a binary mask image which contains change and no change area. SVM or GLC classifier are selected to classify the change area, and get an accurate change information.(4) Change detection software for the geographic national conditions monitoring. On the basis of the geographic national conditions and land cover interpretation system developed by the CASM, integrate the proposed process of multi-level decomposition fuzzy judgment change detection method and optimize high-resolution remote sensing image land cover interpretation system which based on geographic national conditions monitoring.