目的 为了解决当前图像修复算法在破损面积较大时，其复原图像易丢失局部细节信息而导致修复图像存在振铃效应以及不连贯效应的不足，提出一种基于相似稀疏度耦合局部差异特征的图像修复算法。方法 首先，利用待修复块及其相邻块内像素的均方距离来构造相似稀疏度模型，以形成优先权度量函数，根据其计算的优先级来确定优先修复块。然后，通过样本块对应的梯度向量模值来构造局部差异因子，计算样本块的局部差异，并以计算结果为依据对样本块的尺寸进行调整。最后，以像素点的颜色差值信息为依据，构造近似函数，选取最优匹配块对待修复块进行复原。结果 仿真实验结果显示，与当前图像修复算法相比，该算法具有更高的修复质量和效率，其复原图像不存在振铃效应和不连续效应等。结论 所提算法具有较高的修复视觉质量，能用于大面积损坏图像的复原。

The work aims to propose an image inpainting algorithm based on similarity sparsity coupled with local difference feature, to solve the problem that the restored image of the current image inpainting algorithm easily loses the local details when the damaged area is large, thus causing such deficiencies as the ringing effect and the incoherence effect of the restored image. Firstly, the similarity sparsity model was constructed by the mean square distance of the pixels in the block to be restored and its adjacent blocks, so as to form the priority measurement function, and the priority repair block was determined according to the priority calculated based on the said function. Then, the local difference factor was constructed by the gradient vector value corresponding to the sample block. The local difference of the sample block was calculated, and the size of the sample block was adjusted on the basis of the calculation results. Finally, the approximate function was constructed based on the color difference information of pixels, and the best matching blocks were selected to repair the blocks to be restored. The simulation results showed that, compared with the current image restoration algorithms, the proposed algorithm had higher restoration quality and efficiency. The restored image was not subject to ringing effect and incoherence effect, etc. The proposed algorithm has higher visual restoration quality and can be used for the restoration of extensively damaged image.