结合多分辨率与块分割的不同聚焦图像融合

Multi-focus Image Fusion Using Multiresolution and Block Segment

提出了一种结合多分辨率与图像块分割的图像融合新方法。该方法除将多分辨率融合图像块作为可能位于源图像的模糊区域与清晰区域交界处位置的最终融合图像块之外,还将多分辨率图像融合方法获取的融合图像与源图像进行分块比较,选用与多分辨率融合图像块相似的源图像块作为最终融合图像块。实验结果表明,该方法能有效提高常用的多分辨率图像融合方法的融合效果。

A novel image fusion algorithm based on multiresolution and image block segment is proposed, At first the initial fused image is acquired with one of the conventional multiresolution image fusion methods. The initial fused image and the original images are divided into some image blocks with the same size. The image blocks of the original images, which are more similar to the corresponding initial fused image blocks, are selected as the final fused image blocks, while the initial fused image blocks that may be between the clear part and the blurring one of the original images are retained as the final fused image blocks. Experimental results show that the proposed method can improve the quality of the fused images that are obtained with the conventional multiresolution image fusion approaches efficiently.