Research On Multi Source Remote Sensing Image Fusion Based On IHS Transform And Sparse Representation

At present, with the pace of development of space technology change rapidly like, satellite remote sensing data of remote sensing satellite carrying a variety of sensors has been constantly in space for us to send a variety of different space resolution, which makes the research of how to better use of multi-source remote sensing image fusion, image fusion quality in order to obtain better, become a key problem need to be solved urgently in current remote sensing application field. Multi source remote sensing image fusion is to carry out processing of satellite image data from different sensors in the same area, so as to improve the quality of access to information, to achieve information superiority complementation. In this paper, by the fusion of high resolution panchromatic image and a low resolution multi spectral image as the research object, the IHS transform, the classical discrete wavelet transform and the theory of compressed sensing fusion algorithm. According to the IHS transform algorithm caused serious loss of spectral characteristics in remarkably improve the image spatial resolution at the same time, the algorithm analyses the reasons causing color distortion, proposes an improved algorithm.This paper is based on the traditional image fusion framework,-- sparsity structure combined with the wide application of signal processing and its application in the field, puts forward a method of image fusion based on HIS transform and sparse representation, get the remote sensing image by IHS transform I component, H component and S component. Image transform, sparse sparse dictionary based on sparse coefficient corresponding representation of I component and high resolution panchromatic image, sparse coefficient of the two, respectively, using different fusion rules for fusion, OMP fusion algorithm of sparse signal reconstruction coefficients are obtained by the new I component, H component and IHS inverter with the S component of the image fusion result is obtained in the final image. We through the experiment simulation of multi groups of remote sensing image fusion algorithm, compared with the traditional, verify the superiority to improve the algorithm of the.This paper has pointed out the deficiencies in the analysis of IHS transform algorithm, discarded directly I component framework of the original algorithm, ignoring the spectral information contained in the I component. We are as far as possible to keep the I component in the improved algorithm ideas, so as to effectively preserve the spectral information, multiple fusion idea is to use a set of complementary pixel level image fusion algorithm for multiple fusions of I component and panchromatic images based on pixel selection algorithm, used in the new I component. This algorithm improvement in subjective and objective evaluation indexes are showing better advantage.