基于多参数正则化的高分辨率图像重构算法

High Resolution Image Reconstruction via Multi-parameter Regularization Method

根据完全重构滤波器的性质,提出一种基于多参数正则化方法的高分辨率图像重构算法。该算法利用多参数正则化方法,恢复图像的高频部分,然后将恢复的高频信号与低频部分叠加,得到重构的高分辨率图像。由于该算法沿不同滤波方向选取不同的正则化参数,因此有效地抑制了噪声对重构图像的影响,改善了图像在边缘部分的重构效果,从而弥补了己有方法的不足。对所提出算法的收敛性进行了分析,并通过数值实验验证文中所提出算法的有效性。

A high resolution image reconstruction via multi-parameter regularizqtion method is proposed by using on the properties of PRFB. In this algorithm, a multi-parameter regularizqtion metho is adopted to rexomstruct the hith frequency parts of burr images, and a high resolution image is reconstructed by adding the reconstructed high frequency parts with the low frequency part. Since different parameters are set along defferent directions, the effects of noise are controlled, and the reconstructed high resolution image is improved at the same time. The convergence of this algorithm is proved. Experiments show that the proposed methid is effctive.