基于改进复扩散耦合非局部均值滤波器的图像放大

摘要

为了提高偏微分方程放大算法对纹理细节的放大效果，利用改进的复扩散模型耦合均值滤波器，提出了一种图像放大算法。改进的非线性复扩散模型能够很好的定位图像边缘，通过冲激滤波器对边缘进行锐化，同时耦合非局部均值滤波器，保持图像内部的自相似特性，利用非局部信息重建高分辨率图像，提高小边缘、细节放大效果。算法结合非局部信息和局部信息对图像进行放大，增强纹理细节，使图像更加自然，同时减弱对边缘的过度增强，具有较好的放大效果，仿真实验验证了算法的优良性能。%In order to improve the zoomed effect of the method based on partial differential equation,combi-ning the improved complex diffusion model and the nonlocal means filter,the image enlargement method is pro-posed.Having the advantage of precision location of the edges,with the edges sharpened by the shock filter, the improved anisotropic complex diffusion couples to the nonlocal means filter to keep with the similarity of the diffused image,the high resolution image is reconstructed.Not only using the local information of the image but the image’s nonlocal information,the method makes the zoomed image more natural,at the same time,attenu-ates the edge’s over-sharpen,therefore the enlarged image has better visual effective.The simulations prove the prominent performance of the proposed method.