中文摘要：

提出了一种有效的基于局部对比度的分块压缩感知多聚焦图像融合算法。首先采用结构随机矩阵对源图像进行分块压缩测量,获得分块压缩测量值;其次,根据块局部对比度选择清晰的块进行初步融合;再通过多数滤波对初步融合结果进行一致性校验,得到最终的融合结果;最后,通过平滑投影Landweber算法（SPL）重构融合图像。实验结果表明,与目前基于BCS图像融合方法相比,本文所提方法对于多聚焦图像融合,在主观视觉感知以及客观定量指标如信息熵、互信息及平均梯度及算法运行效率等方面均有明显改进。

英文摘要：

An efficient local contrast and block compressed sensing（ BCS） based on multi-focus image fusion algorithm is proposed. Firstly,structural random matrix is used as measurement matrix to obtain a high efficiency sample performance. Secondly,a local contrast measurement in CS domain is proposed to classify the clarity block and the de-focus block,and upon which the larger local contrast block is selected as the fused block. Thirdly,a consistency verification process based on majority filter is introduced to modify the initial fusion CS image. Finally,smoothed projection Landweber（ SPL） algorithm is used to reconstruct the fused image to overcome the block artifact. The experimental results show that,compare to the current BCS based image fusion methods,the proposed method achieves good improvement in subjective visual perception quality as well as in objective quantified quality index such as information entropy,mutual information and average gradient for multi-focus image fusion.