中文摘要：

针对多聚焦图像融合后清晰像素信息损失较为严重的问题,提出一种分层模型与局部复原相结合的方法。首先,根据源图像模糊区域分布情况,对不同灰度特征源图像进行多尺度分解选取恰当的分层模型,并利用此模型对图像进行初步融合;在此基础上,通过计算对应区域的功率谱曲线斜率以推导图像中局部模糊区域像素值,提高图像融合算法的细节提取能力,再结合维纳滤波策略对其进行恢复,以实现图像清晰融合。最后,利用3个实例验证了该方法的有效性与可行性。

英文摘要：

Aiming at the problem of serious clear pixel information loss after multi - focus image fusion, a multi-focus image fusion method based on hierarchical block model and partial recovery is proposed. First of all, according to the distribution of the original image blur area, multi-scale decomposition is performed on different gray distribution feature original images, and appropriate hierarchical model is selected. This model is used to conduct primary image fusion. Then, the pixel values of the partial fuzzy areas of the image could be derived through calculating the power spectrum curve slope in corresponding areas to improve the detail extraction capacity of the image fusion algorithm, the Wiener filter strategy is used to restore these areas, and the clear image fusion is achieved. Finally, three examples are used to verify the effectiveness and feasibility of the proposed method.