the traditional phase matching method has a good inhibited effect on high-frequency noise of images; but the phase-matching algorithm will lose effectiveness because of the low amplitude of band-pass output signal, for more complex images, it is more likely to appear the mismatched phenomenon, and the ability to deal with occlusion region is not very strong. In order to solve the problems above more effectively, this paper has proposed an improved phase matching algorithm and introduced a dynamic programming theory, which has solved blocking and mismatching problems in matching process by the combination with wavelet transform. Meanwhile, the algorithm is also applied to carry out the experiment against noise.
展开▼
