Fusion Of Infrared And Visible Images Based On SCM And Multi-scale Geometric Transformation

Infrared and visible image fusion as an important part of image fusion,has an increasingly broad application prospect in various fields of national economy.Infrared images reflect the physical information by recording the thermal radiation intensity of objects.Visible images describe physical information of objects by recording the reflection characteristics of scene spectral information.The two integration can get richer and more comprehensive images.At present,the fused image based on transform domain become a new development trend,in order to fully utilize the complementary information of Infrared and visible images and the fusion image is obtained better.The proposed algorithm is combined Spiking Cortical Model and multi-scale geometric transform domain such as Non-subsampled Contourlet Transform and Non-subsampled Dual-tree Complex Contourlet Transform to fusion infrared and visible images.The image after multi-scale and multi-resolution decomposition,according to the characteristics of infrared and visible images respectively,high and low frequency are respectively fused using different fusion rules.The main contributions of this paper are as follows:1.Fusion of infrared and visible images based on target extraction and SCM in the domain of NSCTFor infrared and visible image fusion,the target is not prominent enough,In this paper,according to the characteristics of the spiking cortical model which can release pulse synchronously and combined with the region growing algorithm to extract infrared,the infrared image after the target is extracted and original visible image are decomposed by NSCT transform and get the target region and background region,according to the high frequency and the low frequency contains different information.The fusion rules for the high frequency of target region,using the edge energy as SCM external incentive to pulse selection and obtain high frequency coefficients of the fused target region,the final fused image by inverse NSCT transform to obtain.The results of simulation experiments showthat the fusion algorithm of infrared and visible image can retain the target information of infrared images and background information of visible images well in this paper and the fusion effect is ideal.2.Fusion of infrared and visible images based on fuzzy logic and SCM in the domain of NSDTCTIn view of the shortcomings of the existing multi-scale geometric transformation,this chapter proposes to decompose the fusion rule of images by combining the multi-scale,multi-resolution NSDTCT transform with fuzzy logic and SCM.The NSDTCT transform has better characteristics of directional selectivity,translation invariance,anisotropy and time-frequency localization,which can solve the aliasing phenomenon and improve the gradation and clarity of the image effectively.Fuzzy logic is good at expressing the problem of blurred boundaries,which helps to enhance the contrast of the image and is very suitable for the fusion of infrared and visible image.SCM has the advantages both of PCNN and ICM,which is easy to operate and robust to geometric changes,so as to improve the quality of fused images.In this chapter,we first use S function to improve the contrast of the original infrared image.Next,using NSDTCT transform to decompose the enhanced infrared image and the original visible image,the low frequency and high frequency subband coefficients are obtained.Then,in the low frequency,the weight coefficients of the fuzzy function are obtained by calculating local region energy,and then the fused low frequency coefficients are obtained;in high frequency,the fused high frequency coefficients after stimulating the SCM by the pixel clarity and significant.Finally,the fusion image is reconstructed by inverse NSDTCT transform.Compared with the existing mainstream fusion methods,the results show that the algorithm is reliable and effective.