基于超分辨力图像复原算法的模糊系统辨识

由于利用已有的知识和经验得到的点扩散函数(PSF)估计值与真实值之间的偏差会直接影响图像复原质量。为准确地估计PSF，使其向真实的点扩散函数类型和参量逼近，采用超分辨力图像复原MPML算法同几种常用的数字图像去模糊处理进行比较分析，通过实验表明：MPML算法具有其他几种算法的优势，同时减少了对原有信息的丢失；在此基础的同时，按照点扩散函数的分类，分析了点扩散函数的不同估计值及其对图像复原的影响，提出基于超分辨力图像复原算法的图像细节评价参数D，保证了复原图像主观效果和评价参数的一致性。对于模糊图像的系统辨识及其图像复原问题的解决具有实际意义，但对于附加较大噪声条件下的图像复原仍是需要进一步研究的问题。

Identification of Blurred System Based on Super-resolution Reconstructing Image Schemes

For identifying unknown PSF in blurred image and getting more precise reconstructing images,the comparison of super-resolution reconstruction image algorithm MPML with several common deconvolution algorithms was introduced.According to simulation experiments,MPML has the advantages of other,methods as well as decrease losing from original datum.The sorts and various estimating parameters of Point Spread Function(PSF) along with the influence on degradation image reconstruction were analyzed,and the consistency between image-reconstruction subjective effect and evaluatin parameter was guaranteed with satisfaction after one novel evaluating parameter D proposed based on MPML.It is certain of reference to reconstructing blurred images by accurately identifying unknown PSF from blurred images,but still it need more researches on reconstructing field under bigger noise.