Image Enhancement And Fusion Method Based On Nonsubsampled Contourlet Transform

Sensor imaging technology as one part of the field of sensor applications, which isused in many fields, such as satellite reconnaissance, mineral exploration, targettracking, object recognition. Sensor image fusion can describe the target or sceneaccurately, comprehensive from different levels. Sometimes it is hard to l distinguishthe target that we want from the image as a result of the limitation of the weather andreal conditions. So the image enhancement algorithm is used to enhance the target weinterested in order to highlight the target we are interested.In this paper, to solve the statements problem of the contour information inwavelet transform、pseudo-Gibbs phenomena around singularities and the poor in visual effect, new image enhancement and fusion method are proposed which baseson the nonsubsampled contourlet transform by the analysis the methods in imagefusion and enhancement domain recently. It consists of nonsubsampled pyramid andnonsubsampled directional filter banks. It has the characteristic of rich directionalinformation and shift invariance, which can describe texture information and solve thepseudo-Gibbs phenomena around singularity in the image decomposition process. Atthe same time, the pulse coupled neural network, fuzzy contrast, the energy of edgeinformation, HIS, PCA transform combined with the nonsubsampled contourlettransform which has the characteristic of multi-directional, shift invariance,anisotropicare introduced into the proposed method to deal with the selectedexperimental image in nonsubsampled contourlet domain. The proposed methods areused in remote image enhancement, multi-focus image fusion, multi-spectral imageand panchromatic image fusion which get good visual effort，improving the contrast,definition and so on by inspecting of the evaluation index.