Study on the whole seam-line removal of remote sensing image based on class discrimination

Image mosaicking is a very important approach in the image processing, it makes multiple digital image makes up a large image of the complete. The two(multiple) images which are mosaicked have quite differences in the whole gray(color), and this will inevitably produce consistent color image mosaic after mosaic, so it is necessary to use the whole seam-line removal. Unlike conventional we just do not take into account differences in the whole color, simply processing of the mosaic seam-line, in this paper, the using of class discrimination targeted to achieve the whole seam-line studies provide the theoretical support for the application of remote sensing image.