针对惯性约束聚变(ICF)中环形孔径编码图像的恢复问题提出了一种改进的维纳滤波方法。在传统的维纳滤波方法中,由于原图像和噪声是未知的,故通常是用某一待定常量来代替其中的噪声与信号的谱密度之比。这种近似忽视了信号与噪声本身的信息,从而造成丢失某些关键的细节,难以达到高质量的图像复原效果。在改进的方法中,首先采用传统维纳滤波方法求得初始估值,然后利用该初始值求得原图像及噪声的谱密度估值,进而利用这些新获得的信息构成改进的维纳滤波器对退化图像进行第二次滤波。实验表明,这种改进方法可以克服原方法的不足,突出图像的一些关键细节,提高图像的整体质量。在仿真实验中,恢复图像的均方误差降低了15%以上;在实际惯性约束聚变图像的解码恢复实验中,图像恢复效果亦有显著改善。该方法还可以推广到其他图像恢复的应用中。

A modified Wiener filtering method for image restoration of the ring-coded aperture images in inertial confinement fusion (ICF) is presented. In traditional Wiener filtering method, the signal-to-noise ratio of spectral density has to be replaced with a constant parameter, which overlooks the spectral information of the underlying image and noise because the spectra are a priori unknown, and consequently the filtered result may lose some important details with inadequate resolution. In order to overcome this drawback, the authors propose an approach to exploit the spectral information. In the new method, the traditional Wiener filter is used at first to obtain the initial estimate of the target image, based on which the spectral densities of the target image and noise image are estimated. Then by using this information a new Wiener filter is reconstructed that is used to filter the coded image again to obtain an improved result. The validity of the new method is confirmed by the experimental results