摘要
图象的盲解卷积恢复具有重要的理论和实际意义,许多情况下系统的扩散特性不能精确获得。针对一类相对平滑或类似高斯分布的扩散特性,建立一种图象盲解卷积算法,采用交替迭代方法,适合总体最小二乘求解。算法能有效地确定点扩散函数,图象恢复质量有明显改善。最后的仿真实验表明了算法的有效性和稳定性。
In many image restoration applications, the point spread function(PSF) is not known exactly. We establish a blind deconvolution algorithm for images, where the PSF is smooth or similar to that of Gaussian blur. Using alternative iterative steps to solve the minimization problem, our algorithm is suitable for the total least square (TLS) method. This algorithm is efficient because different PSFs can be identified at the same SNR without changing parameters. Final experiments are presented to demonstrate the effectiveness and robustness of the algorithm.
出处
《计算机工程与科学》
CSCD
2004年第4期42-44,94,共4页
Computer Engineering & Science