采用噪声自适应正则化的图像重建的仿真研究

针对图像分辨率过低的情况,传统的全局正则化方法未能有效地分辨图像的细节,尤其是图像的边缘问题,根据图像的降质模型,基于正则化方法,并结合降质图像模型,采用迭代的方式确定局部正则化参数代替全局化的正则化参数,提出一种基于噪声的自适应正则化方法的图像重建.再次通过Matlab软件仿真,对比全局正则化和新的基于噪声自适应正则化图像重建的结果,仿真实验证明该方法能够克服传统的正则化方法的局限性,并能够很好地复原图像的边缘细节.

Simulation Rearch of Image Reconstruction Using Noise Adaptive Regularization Method

In order to improve higher resolution,traditional global regularization method can＇t solve how to identify image detail ,especially the border of image. This text according to the image degradation model, based on the regularization method, and combined with the degraded image model, using local regularization parameter by iteration method instead of global parameter, it gives an adaptive regularization method for image reconstruction based on the noise. And then with the Matlab software simulation,this palper compares global regularization with improved adaptive regularization image reconstruction based on noise,and it is proved new method can overcome the shortcomings of traditional regularization method limitation, and can well recover the image edge details.