离焦模糊图像的维纳滤波复原研究

微操作中，显微视觉系统获取的图像通常是离焦模糊图像。离焦模糊图像的退化模型可用圆盘函数描述，利用模糊图像无方向性的二阶拉氏微分图像的自相关的负相关峰形成的环形槽的直径等于作为圆盘函数直径的2倍可以确定该函数。对模糊图像进行一次维纳滤波方法得到原图像的估计值，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，该方法计算量小、鉴别精度高、抗噪声能力较强，突出原图像的一些关键细节，提高了图像的复原质量。

Study on Wiener filtering for restoration of defocus blur image

Images acquired by microvision system are usually defocus blur images in micromanipulation. The defocus image can be depicted by uniformly distributed function on circular support region. Firstly, the Laplacian operator that is an isotropic second order differential arithmetic operator is employed to derive the defocus image, and the autocorrelation of the derived image is calculated by FFT. The blur parameter can be estimated as a half of the radius of the annular groove formed by the minus peaks of the autocorrelation image. Wiener filtering method is used first time to obtain the initial estimate of the target image, then, based on the initial estimate the spectral densities of the target image and noise image are estimated. Using this information, the Wiener filter is reconstructed and used to filter the defocus image again to obtain an improved result. The method has the advantages of high precision, less calculation task, and strong anti-interference capability. Some experiments were performed to validate the performance of this method, and experimental results show that the image restoration results are improved significantly compared with the results obtained by using traditional Wiener filters.