基于对比度视觉模型的多聚焦融合图像的精确重构

Accurate Reconstruction of Multi-Focus Fused Image-Based Contrast Vision Model

采用基于对比度视觉模型的图像融合最优分块搜索算法 ,对同一场景两幅严格配准的多聚焦图像的清晰恢复进行了深入研究 .结果表明 :应用二十多幅各种不同图像实例 ,该方法的评价指标 RMSE(均方根误差 )不仅远小于目前国内外文献中研究的数值 ,甚至在一些情况下达到 0 ,实现了多聚焦图像对原始图像的精确重构 .该算法亦具有简捷性和可扩展性 。

Multi-focus image fusion is to combine information from multiple images of the same scene for producing a merged image. A reconstructed scheme based on contrast vision model was studied briefly exploiting segmentation techniques and blocks search. More than 20 images demonstrate evaluation measure RMSE (root mean square error) is far less than the values found in the literature at present, even 0 to some cases. The fused image achieves accurate reconstruction to the original image. The optimal fusion algorithm can also be readily extended to the misregistration of the images which can obtain further reappearance.