Multispectral Sensing of Satellite Images for the Classification of Different Land Covering Area by Support Vector Machine-2 Method

The problem of scarcity of multi image supporting for satellite image is handled with the help of multispectral sensing image method. In that the SVM-2 (Support Vector Machine - 2) is consider for classifying the different types of land covering area. The workflow of multispectral acquired with an absurd image into orthorectified image, in that we identified several challenges including file format compatibilities and a size of the original image. These compatibilities are handled with the help of segmentation and semi supervised learning algorithm. The segmentation is used to simplify the portrayal of large image into something more meaningful and easier to analyze. The Multispectral provides a quality and high resolution information for satellite sensor applications. With the help of semi supervised learning algorithm and multispectral sensing image the overall performance of PSNR is increased upto 42.98%.