基于倒频谱的无噪声运动模糊图像快速复原算法

Fast Restoration Algorithm of Noise-Free Images Based on Cepstrum

针对当前运动模糊图像复原算法主要集中在非实时性研究,着重考虑算法的执行质量而不考虑算法执行效率的问题,提出了基于倒频谱的无噪声图像的快速复原算法.该算法对无噪声模糊图像进行倒频谱处理,从倒频谱图中直接得到图像的模糊方向和模糊距离.对于尺寸较大的图像,在执行倒频谱算法前先使用双线性插值算法实现图像压缩,再进行倒频谱操作.通过比较各种复原算法的执行效率及效果,采用维纳滤波对图像进行复原.通过对多幅模糊图像进行仿真复原试验,在获得了良好的复原效果的同时也取得了较高的执行效率.

For the problem of current motion blurred image restoration focusing on non-real-time research,that is,focusing on the quality of the algorithm rather than considering the executed efficiency of the algorithm,a fast restoration algorithm of noise-free images based on cepstrum is raised.The blur direction and blur length are directly generated from the cepstrum,which is elaborated by the algorithm of noise-free image.For images with large size,a compression of bilinear interpolation is completed initially then goes the cepstrum of the compressed image.The wiener filtering for image restoration is selected under the comparison with various restoration algorithm,regarding the efficiency and effectiveness.Hence through the deblur experiments on blurred images,restoration effect can be demonstrated along with the high performance efficiency.