中文摘要：

非下采样轮廓波变换（NSCT）中存在低频子带系数不稀疏、分解级数难以选择的缺点,而稀疏表示的融合方法易造成图像的纹理和边缘等细节趋于平滑。针对上述问题,提出一种基于NSCT和稀疏表示的多聚焦图像融合算法。对图像进行1级NSCT分解;对低频子带系数采用稀疏表示的方法进行融合,对高频子带系数采用一种方向特征对比度取大的方法进行融合;经NSCT逆变换后得到最终的融合图像。实验结果表明,该算法提高了图像的主观视觉质量,在客观评价指标上也有所提高。

英文摘要：

The problems of lower sparseness of low-frequency sub-band coefficients and the difficulty in selecting the NSCT decomposition level exist in the non-subsampled Contourlet transform （NSCT）. And the fine details in source images like textures and edges tend to he smoothed using the sparse representation algorithm. To solve these problems, an algorithm based on NSCT and sparse representation was proposed. The source images were transformed using NSCT. Sparse representation was used to fuse the low-frequency sub-band coefficients, and the high-frequency sub-band coefficients were fused using a method which selected the maximal contrast values of its directional feature. The fusion image was reconstructed using the inverse NSCT. Experimental results demonstrate that the algorithm proposed improves both subjective visual effect and objective evaluation index.