基于各向异性和非线性正则化的湍流退化图像复原

提出了一种基于各向异性和非线性正则化的湍流退化图像复原新算法。为了有效地从湍流退化图像中估计出退化模型，在湍流点扩展函数的优化估计过程中合理地融合了有关湍流点扩展函数的一些基本的先验知识。首先，将点扩展函数的非负性和光滑性约束加入到目标函数中。再针对湍流点扩展函数的衰减性质，建立了一个具有非线性和空间各向异性的正则化函数，使其在重建点扩展函数时能适当地进行空间梯度平滑。由此通过迭代极小化目标函数来优化估计湍流点扩展函数值，进而恢复图像。对各种噪声条件下的湍流退化图像进行了恢复实验，实验结果表明本文算法具有较强的稳定性和抗噪能力。

Restoration of turbulence-degraded images using anisotropic and nonlinear regularization

A new algorithm is proposed for restoring the turbulence-degraded images by use of anisotropic and nonlinear regularization.In order to effectually estimate the turbulence-degraded model from turbulence-degraded images,some a priori knowledge about turbulence point spread functions (PSFs) is incorporated properly in the process of optimal estimation of turbulence PSFs.First,the constraints of non-negative and smoothing are added to the objective function.Second,to match the decay nature of turbulence PSFs,an anisotropic and nonlinear regularization function is suggested to adequately regulate the PSF values in the process of rebuilding PSF.Thus,the PSF values can be estimated by iteratively minimizing the objective function,and the object image can be restored.Experiments results are reported in the restoration of turbulence-degraded images at various degrees of noise,which show that the proposed algorithm is robust with high ability of resisting noise.