Abstract: This paper introduces image repairing model based on improved total variation. And the difference of gradient in image region will bring about different influence on image repairing. This text considers the influence of human vision psychology on image effect. A new variation function is proposed and the Euler-Lagrange equation of the model is put forward. This model proposed by this text is applied to a still image, and the contrast on image restoration effect is made. Still image repairing model is applied to video repairing problems. A frame is added between two adjacent frames, and then objectives are put on corresponding positions. So a new frame is produced and persons’ motions are made consistent, so that better video image repairing effects are achieved.

Key words: improved TV model, video repairing, new frame, vision psychology

摘要： 改进了经典的图像修补变分模型，既考虑了图像区域中梯度不同给图像修补带来的不同影响，同时兼顾了视觉心理学对图像效果的影响。提出了新的变分泛函，给出了相应的欧拉-拉格朗日方程。将此模型应用到静态的图像上，和原来的模型在图像修补效果上进行了对比。将此图像修补模型应用到视频修补问题上，从一个新的角度，在两帧图像之间以增加一帧，把目标放在相应位置，形成新帧，使人物动作连贯，达到更好的视频图像的修补效果。

关键词: 改进的TV模型, 视频修补, 新帧, 视觉心理学