基于PCNN的图像融合新方法

本文提出了一种基于PCNN的新型图像融合算法.首先,对待融合的两幅图像进行平稳小波分解得到两组多尺度图像;接着,取其中任意一组作为主PCNN的输入、另一组相应的图像作为从PCNN的输入,在每次迭代时,经并行PCNN点火后,得到一系列多尺度融合图像;然后,对它们进行平稳小波反变换得到每次迭代的融合结果;最后,计算每次迭代结果的信息熵,取信息熵值最大的融合图像作为最终结果.大量的实验以及与其它融合算法的比较分析,表明了本文算法的有效性和优越性.

New Image Fusion Algorithm Based on PCNN

A novel algorithm based on Pulse Coupled Neural Network (PCNN) for image fusion was proposed.First,the two original images were decomposed by stationary wavelet transform,meanwhile,the two group multiscale sequences of each input images could be obtained.Secondly,one of the multiscale sequences were chosen arbitrarily as the input to the main PCNN network,and the others as the input to the subsidiary network.Then,sequences of multiscale fusion images were gotten by the parallel PCNN and the fused image could be obtained by inverse stationary wavelet transform at each process of iteration.At last,the optimal fusion result is obtained when the maximum value of the information entropy is achieved.Lots of experiments and comparisons with other fusion algorithms show the effectiveness and superiority of new method.