图像修复是数字图像处理的重要内容, 可用于被损坏的图像和视频修复、视频文字去除以及视频错误隐藏等。基于tv模型的修复方法有较好的恢复效果, 但对参数的选取比较敏感, 运算量较大。本文提出了一种基于tv模型的自适应图像修复方法, 与原方法相比可以有效提高该算法的稳健性, 并能显著的减少运算时间。-Image inpainting is the important content of digital image processing, can be used to repair damaged images and videos, remove text and video video error concealment. Based on the TV model repair method has better recovery effect, but the parameters selection of sensitive, computation. This paper presents a TV model based adaptive image restoration method, compared with the original method can effectively improve the robustness of the algorithm, and can significantly reduce the computational time.