边缘振荡效应的自适应滤除算法

超分辨率图像重构技术已经被证明在卫星成像、视频监控、医学成像等研究领域中具有极其重要的应用价值。但是,由于超分辨率图像重构本身是一个病态求逆过程,在重构图像中所产生的边缘振荡效应往往很难在重构过程中予以消除。为此,文章提出了一个边缘振荡效应的自适应滤除算法。在该算法中,重构图像先被分割为均匀区域、边界区域和纹理区域,然后分别对这三类区域采用不同的滤波算法。实验结果表明,新算法在滤除边缘振荡效应的同时,较好地保存了超分辨重构图像的细节。

Ringing Artifacts Reduction Using Adaptive Filtering

The Super-resolution technique has been proven to be extremely useful in satellite imaging,video surveillance,medical computed tomographic imaging,and other applications.However,due to the ill-posed nature of the super-resolution image reconstruction,the ringing artifacts can not be easily removed in the reconstruction process.In this paper,an adaptive ringing artifacts reduction algorithm is proposed.In the proposed algorithm,the pixels are first classified into three categories:pixels belonging to homogeneous regions,pixels belonging to edge,and pixels belonging to texture regions.A new adaptive spatial filter is then used for image enhancement.Experimental results demonstrate that the proposed method can suppress ringing artifacts greatly,while the fine details in the super-resolved image are preserved.