Non-downsampling Contourlet Transform-based Image Fusion Algorithm

Image fusion is the technique that integrates complementary and redundant information of multiple images of the same scene to obtain a composite one, which contains more complete and accurate description of the scene than any of the individual image. After the image fusion, credibility and clarity have improved ,and their intelligibility are better for human vision and computer detection, classification, recognition, understanding and other processing.Multi-focus image fusion is an important branch of the image fusion. In a scene imaging, we need focus on different objectives from access to multiple images, as making the same focal lengths can not get clear different images which have clear multiple targets in different depth of scene, because of limited focus of optical imaging system, and than, fusion generates a clear image of various targets by extracting a clear image of their respective regions. At present ,most of fusion algorithms use fusion rule picking low frequency sub-band fusion coefficient by a method called"average way", and that makes some useful information of the source image lost. For the problems we talked above, a multi-focus image fusion algorithm was proposed in this paper which combining local entropy and Nonsubsampled Contourlet Transform. And we do a comparison experiment, the experiment results show that the way significantly improved fusion results.The sensor of Infrared and visible are widely used in the field of military and security monitoring. Visible image is reflected images, which have much high frequency components, it can reflect the details of the scene over a certain light intensity, however, when over poor light intensity, the visible images (ie, low light level image) are have a low contrast; Infrared image is radiation image, whose gray is decided by the temperature difference between target and background, and it does not reflect the real scene. There are shortages on Visible or infrared images which works alone, image fusion can effectively integrate their characteristic information, enhance understanding of the scene, highlight objectives, detect targets quickly and accurately at the situation of hiding, disguising and confusing. This paper gives out a new method, which based on the complementary imformation characteristics of infrared and visible image and nonsubsampled contourlet transform, and experimental results show that the method is effective.