Abstract: After using the motion estimation based on optical flow to make the core image registration，and because of the fuzzy part of rugged core image，a new multi-focus image fusion method based on Curvelet transform is proposed.First，two different focus images are decomposed using Curvelet transform separately，then the new Curvelet coefficients are acquired by adopting a simple fusion rule，i.e.low-frequency coefficients uses the weighted average and high-frequency coefficients uses the largest absolute value.Finally，the fused image is obtained by performing an inverse Curvelet transform.Thus the experiment demonstrates that the fusion method based on Curvelet transform can effectively integrate the multi-focus images.Compared with wavelet transform method，the proposed method can get better fusion effect.

Key words: image registration, Curvelet transform, image fusion, multi-focus image

摘要： 利用基于光流的运动估计对岩心图像配准后，针对岩心高低不平图像出现部分模糊现象，提出了一种基于Curvelet变换的多聚焦图像融合方法，先对不同聚焦图像进行Curvelet变换，采取低频分量取加权平均、高频分量取绝对值最大的融合规则，再进行Curvelet反变换。实验表明，基于Curvelet变换的融合方法可有效综合多聚焦图像，相比基于小波变换的图像融合，能获得更好的融合效果。

关键词: 图像配准, Curvelet变换, 图像融合, 多聚焦图像