中文摘要：

为消除图像消噪中出现的Gibbs现象，基于Eno—Haar模型，提出一种新的图像修复算法。利用小波分解的遗传特征，将间断信息遗传到各分解层中；对带有间断信息的小波系数，设计多层多方向插值算法进行插值，即在各个分解层上分别插值，并考虑多个方向进行加权平均，得到最优的小波系数，重构得到最优的修复图像。最后对算法进行仿真实验，实验表明此算法具有很高的可行性和应用推广价值。

英文摘要：

A new image recovery algorithm based the Eno - Haar model is devised in order to eliminate the Gibbs phenomena generated in image denoising. Transfering the discontinuous information to the wavelet coefficients by making use of the genetic between the wavelet coefficients. Afterward,modifying the coefficients by multi- level and multi-orientation interpolation algorithm, and optimizing the interpolation results by weighted average method to obtain the optimum coefficients, the recovery images are acquired through the wavelet reconstruction. Finally, the algorithm is verified to be feasible and efficient by simulation experiments.