分析了基于整体变分（total variation,TV）模型的图像修复算法,TV模型修复算法只使用各向异性扩散,TV模型各向异性扩散仅向图像边缘方向扩散,容易在平滑区域引入阶梯效应。提出了一种改进的图像修复算法,该算法同时结合了各向同性和各向异性扩散,利用区域频率差异实现了在不同的区域使用不同的迭代方程,有效避免了原始算法引入的阶梯效应,同时在平滑区域提高了迭代效率。Matlab环境下的仿真结果表明,改进算法的修复效果和峰值信噪比的计算结果均明显优于原始算法-Analysis based on the total variation (total variation, TV) image restoration algorithm model behavior, TV model repair algorithm uses only anisotropic diffusion, TV model anisotropic diffusion diffusion only to the image edge direction, easy to introduce step effect in smooth areas. An improved image restoration algorithm, which combine both isotropic and anisotropic diffusion, use of regional difference in frequency enables the use of iterative equations different in different regions, effectively avoiding the staircase effect of the original algorithm introduced while smooth regions to improve iterative efficiency. Matlab environment simulation results show that the improved repair effect of the algorithm and the results were significantly better than the peak signal to noise ratio of the original algorithm