Research on fusion algorithm based on wavelet transformation in SAR image identification

Abstract

Aimed at the puzzle existed in each kind of image fusion algorithm in SAR image identification, such as twisted in spectral characteristic, high in algorithm complexity, large in internal memory demand, difficult to give consideration to take account of monolithic construction and detail of image and so on, the paper proposed a sort of image fusion algorithm based wavelet transformation. The algorithm carried respectively through wavelet decomposition for fused original image according to orthogonal wavelet transformation. The low frequency coefficient matrix of fused image wavelet was made of low frequency coefficient was selected by wavelet coefficient in terms of proportion. The high frequency coefficient was used to analyze the edge characteristic of different high and low frequency coefficient in special area by means of texture consistency measure principle, and the coefficient matrix of high frequency wavelet of fused image was determined according to certain rule. Then the algorithm could be formed to carry through the fusion processing for image to obtain the fusion image. The fusion image experiment results of SAR showed that it not only could distinguish forge edge better, make image detail richer and truer, but also could take image whole structure and detail into consideration, and therefore it is higher in identification efficiency.