中文摘要：

通过分析代数多重网格（algebraic multi-grid,AMG）算法中粗网格提取过程,提出了一种基于代数多重网格算法的图像重构算法。在代数多重网格算法的粗网格序列中,下一层粗网格保留上一层网格的强连接部分。将这种机制运用到图像,提取的粗网格可以较好的保留图像的有效信息部分,在图像变化剧烈的细节区域网格点分布不均匀,平滑模糊部分网格点分布均匀一致。以粗网格像素点进行插值,可以得到较好的重建结果。以均方误差为评价参数,与小波算法进行了比较,比较结果表明该算法在一定程度上优于传统的小波算法,且有一个图像融合应用实例,优于小波融合方法。

英文摘要：

After analyzing coarse mesh extraction process in the method of algebraic multi-grid（AMG）,a new way of image reconstruction based on algebraic multi-grid algorithm is proposed.In the coarse mesh sequence of algebraic multigrid algorithm,the next layer of coarse mesh retained the strong connection part of points in the former layer.This mechanism is applied to the image,then the coarse grid extracted from the image can retain the important part of image pixels better,and in the region of coarse mesh where the image changes dramatically change,the distribution of points is uneven while the smooth fuzzy part of the grid point distribution is uniform.After the image interpolation with the coarse grid pixels,a better reconstruction result could be gained.Based on the mean square error evaluation parameters,compared with the wavelet algorithm,the experiment is verified that this algorithm is better than the traditional wavelet algorithm to some extent,and the proposed algorithm is applied in image fusion,better than fusion result by the way of wavelet.