Research On The Application Of Image Fusion And Second-generation Bandelet Transform In Image Denoising

In our daily live, we have been exposed to the image information directly or indirectly. The image is rich of information, it can not only bring people the intuitive information, texture information, and geometric shapes, but also can be used for hiding confidential information. High-quality picture is bound to contain a rich amount of information, however, we collect images, they are more or less mixed noise. Therefore, this thesis will focus on the image denoising, and do the research about the Bandelet Transform Ⅱ and image fusion through.First of all, this thesis gives the description about the research background of image processing, which emphasize the issue of image denoising. The filtering in the space domain and transform domain is the main method for process of image denoising. The final choice of the second generation Bandelet transform with image denoising, has been discussed in this thesis. Meanwhile, this thesis gives the introduction of the threshold function for the threshold denoising. Through the comparison of the adaptive threshold, the combinational method of BayesShrink threshold and Bandelet Transform Ⅱ has been proposed in order to get the more effective denoising result.Then, learning processes and principles of image fusion technology. This thesis also gives out several kinds of the decisions of image fusion and the evaluation criterions of image fusion in detail. Based on the DWT image fusion, and with the effective denoising algorithm of Bandelet Transform adaptive threshold, the denoising algorithm based on both adaptive Bandelet Transform Ⅱ and image fusion technique has been produced. The RMSE evaluation criterion has proved the advantage of the above denoising algorithm.Finally, a recapitulative summary about the research of this thesis, some practical problem, improvement method and the future of image denoising have been concluded.