一种基于Curvelet变换的图像融合新算法

A new image fusion algorithm based on curvelet transform

基于Curvelet变换提出了一种新的融合算法。首先将图像进行Curvelet变换,然后对粗尺度系数和细尺度系数采用不同的融合规则将Curvelet系数融合,最后进行重构得到融合结果。对得到的融合图像进行了主客观评价和对比,实验结果表明,本文方法得到的融合图像,在熵、均方误差(MSE)、峰值信噪比(PSNR)和相似度等客观指标上都优于其他方法。

A new image fusion algorithm based on Curvelet transform is proposed.Firstly,the source images are decomposed by using Curvelet transform,then the coarse scale coefficients and fine scales coefficients are fused with different fusion regulars,and finally the fused coefficients are reconstructed to obtain the fusion result.The obtained fusion image is evaluated in subjective and objective factors.Experiments show that this method can yield fusion images with better visual quality and better objective evaluation criteria in entropy,mean square error,PSNR,similarity and so on.