研究了周期边界条件下，Tikhonov正则化的固定点算法，提出了变化正则化参数的方法。首先对正则化参数取较大值，抑制复原图像中的噪声，通过得出的收敛结果来修正初始梯度; 然后对正则化参数取较小值，以增强复原图像中的细节。实验结果表明，与当前求解L1范数正则化函数和全变分正则化函数的流行算法比较，本文算法对于运动模糊与高斯模糊图像的复原效果更佳。

We analyze the fixed point method with Tikhonov regularization under the periodic boundary conditions, and propose a changable regularization parameter method. Firstly, we choose a bigger one to restrain the noise in the reconstructed image, and get a convergent result to modify the initial gradient. Secondly, we choose a smaller one to increase the details in the image. Experimental results show that compared with other popular algorithms which solve the L1 norm regularization function and Total Variation(TV) regularization function, the improved fixed point method performs favorably in solving the problem of the motion degradation and Gaussian degradation.