Abstract: The acquisition of the point spread function is the key point in non-blind restoration algorithms for image, it’s very important to get the estimation of point spread function parameter. Since the Gaussian function is the most common model in many optical systems and imaging systems, it’s always the hotspot to estimate the parameters for Gaussian point spread function. This paper gives a detailed introduction of principle for three main estimation algorithms for point spread function, and also gives a comparison of the advantage and deficiency among these algorithms. Finally, the paper gives a perspective of the trend for future research.

Key words: polynomial transform, point spread function, blur estimation, wavelet transform, lipschitz exponent

摘要： 在图像的非盲复原算法中，点扩展函数的获取是算法的关键，因而点扩展函数参数估计是图像复原中重要的一步。由于高斯型点扩展函数是许多光学系统和成像模型中最常见的，所以高斯型点扩展函数参数估计算法一直是研究的热点。文中详细论述了三种主要的高斯型点扩展函数参数的估计算法的原理和实现，比较了各种算法的优点和不足之处。最后，文中展望了今后研究的方向，具有重要的实际意义。

关键词: 多项式变换, 点扩展函数, 模糊估计, 小波变换, 李氏指数