Research On Pixel Level Fusion Algorithm For Multi-Source Remote Sensing Image Fusion

As one of the important aspect of image fusion, multi-sensor remote sensing image fusion has attracted many attentions in remote sensing field. The multi-source remote sensing image fusion technology carry the multiple imaging information on the pretreatment. The information is works in the different wavelength range, with a different imaging mechanism of the image sensor, and carry on to the identical scene. Then the technology uses certain algorithm to fuse various images to a new image. So the new image has higher credibility, few fuzzy, better intelligibility. It more suitable for human visual and computer examination, classification, recognition, understanding and other treatment.This paper firstly introduced the development status of the multi-source remote sensing image fusion and the basic theory of image fusion. It also analyzed the pretreatment work which before image fusion. Then from the space domain and transform domain introduced methods of traditional remote sensing image fusion. It also analyzed the pros the cons and the applicable scope of each method.Secondly, in present commonly used image fusion technology, the image fusion algorithm based on multi-resolution analysis is widely used and it is an extremely important algorithm. This paper has done the detailed research to the nature and the principle of the Contourlet transform and the nonsubsampled Contourlet transform, described the remote sensing image fusion algorithm based on the nonsubsampled Contourlet transform. It used a fusion rule based on region selection, and through experiments demonstrate the feasibility of the algorithm.Then this paper investigated the subjective and objective evaluation standard, and analyzed the image across-the-aboard in the experiment by adopting the evaluation criteria combining qualitative analysis with quantitative analysis. By analyzing the experiment results, it also summed up the characteristic and circumstance in point of some common fusion algorithm, approved that this algorithm was better than traditional fusion algorithms, and provided more scientific basis for the application of remote sensing image fusion.At last, this paper summarized the investigative fruits and brought forward fancies and prospects for the problem and unfinished work.