Research On SAR And Panchromatic Remote Sensing Image Fusion In Wavelet Domain

As the current most widely used types of the remote sensing images, SAR and panchromatic remote sensing (PRS) images have different performance on the surface features of the same target owing to their imaging mechanism, such as contour, texture and color information. Based on complementary of SAR and PRS images, a new fused image from them can eliminate the uncertainty of the images. As a result, the fused image is more helpful for further image interpretation, target identification and decision.In this paper, Shift-invariant Discrete Wavelet Transform(SIDWT) and Bandelet transform algorithm are used for fusing SAR and PRS images in pixel level to enhance the complementarity and details of remote sensing images. The main research contents of this paper are as follows:Firstly, the paper studies the theory related to image fusion based on wavelet domain. After the research of wavelet transform, the general procedure of image fusion algorithm in wavelet domain is given in this paper. Then pretreatment for the fusion of SAR and PRS images and the evaluation criteria of image fusion results are discussed.Secondly, a improved SIDWT algorithm with new fusion rule is proposed. Aiming at the weakness of shift variant of the traditional wavelet transform, the SIDWT algorithm and a new rule are proposed to fuse SAR and PRS images. The experimental results for real images demonstrate the effectiveness of the algorithm proposed in this paper. And the algorithm is applied to solve the practical problem of intertidal zone extraction, which has a practical significance.Last but not least, this paper presented a image fuse method based on Bandelet, in which three rules for selecting the geometric flow of SAR and PRS images are proposed. Taking into account the geometric characteristics of SAR and PRS images, such as the direction and continuity of the edges and contours, Bandelet transform method is applied for fusing SAR and PRS images, which takes full advantage of the image geometry Regular features and image information. In the fusion process, three rules for selecting the geometric flow of SAR and PRS images are proposed. The experimental results indicate that the method can give out better fusion performance than wavelet transform.