在空间域内实现运动模糊图象的恢复

Restoration of Motion Blurred Images in Spacial Domain

分析了因水平直线匀速运动而引起的图象模糊,提出在图象的空间域内,直接利用含约束条件的最小二乘法,求解复原图象。本文着重演算了如何利用一维卷积及约束条件来构建代数方程,并给出了复原图象的求解公式。该算法已通过MATLAB编程实现。实验结果表明,与频域算法相比,该算法简单并能得到较好的复原图象,并且无需事先知道噪声模型。

This paper analyzed blurred images caused by the unidirectional constant velocity, and suggested one solution of the restored image in the spatial domain by using the constrained least square criterion. It demonstrated how to derive the algebraic expression of restored images step by step using the convolution expression and the constrained condition. The algorithm was coded in MATLAB and applied on a typical case. The experimental results show that it can restore images with good quality and no prior knowledge on the noise model is required compared with the algorithms based on frequency domain.