Research On Heterogeneous Multisensor Image Fusion Technology

Heterogeneous multisensor image fusion is a process which fuses two or some different species sensor images to get the fusion image. Because of the complementary in visible image and infrared image, there are tremendous application value in military and civil fields. This paper selects visible image and infrered image to fuse and focuses on the fusion algorithm.Frist, based on studying the relevant theories of image fusion and emphatically analysising the registration of image preprocessing, the two methods’ image registration which are Fourier-Melin transform and SURF have been accomplished.Second, through formulating the fusion rule, this paper get three differert fusion algorithms’ fusion images based on wavelet transform and NSCT. Through Matlab experiment, the paper get the fusion images and objective evaluation index.Last, the paper put forward its own fusion algorithm. The low frequency and bandpass subband coefficients of original ima ge are obtained by NSCT. Through formulating the fusion rule, take the decomposed subband coefficients as PCNN’s input to get the fire mapping images. The method of maximum absolute value and region standard deviation is used to choose the fusion coefficie nts of fire mapping image in low frequency part. For bandpass subband coefficients’ fire mapping image, the method based on maximum absolute value and region energy is adopted for the fusion coefficients in the bandpass part. Finally, the fused image can be obtained by inverse transform of NSCT, and get the fusion image’s objective evaluation index through matlab. Compared with the above three algorithms, experiment shows that the paper’s method can get the effective and superior result.