Based On Spot5 Remote Sensing Image Automatic Classification And Treatment Research

The rapid devel pment of geospatial and information technology play an important role in the industry, remote sensing image data is very important as the underlying data, SPOT5 remote sensing image data as a data type in which a variety of planning and design, investigation and analysis and evaluation of land resources has played an important role. SPOT5 based on remote sensing image, focusing on analysis of remote sensing image classification method and the classification of post-processing method.The main content of paper in the following areas:(1) First introduced the background paper, a brief introduction and application of remote sensing images SPOT5, then analyzes the importance of remote sensing image classification; (2) Describes the basics of remote sensing images SPOT5, listed on the pretreatment SPOT5 remote sensing images were corrected for polynomials for the fine images, weighting image fusion, image enhancement based on HIS transformation, some of the major shows this to be SPOT5 classified sources and remote sensing images and image preprocessing; (3) Details the law on supervision of remote sensing image classification and basic principles of the basic idea, from the definition of spatial characteristics of the image to the classification of sample selection, to the maximum likelihood method used to monitor law classification, mainly from the theoretical description and classification ideology and Methods; (4) Act on the supervision of remote sensing image classification using a lot of computer software, paper-based ERDAS remote sensing software, the actual force to be regional or local SPOT5 remote sensing image classification image as an example the various steps and processes, illustrated, introduction of remote sensing image classification preliminary results of image classification accuracy and a simple evaluation and analysis:(5) Supervision of remote sensing image classification method can only be considered as preliminary results of image classification image, in terms of image classification accuracy is measured in terms of visual effects are not ideal, so it is necessary to classify the initial classification of image post-processing, this paper post-processing method used in the main image smoothing, statistical analysis and removal of cluster analysis, in addition to raster image to vector data conversion, through ERDAS classification of post-processing, has been better in the raster image, and been better for GIS vector data analysis; (6) Finally, make a summary of the entire thesis on remote sensing image classification and expectations of future work are described.