【作者】 侯榆青； 张欢； 杨旭朗； 陈燊；

【Author】 HOU Yu-qing,ZHANG Huan,YANG Xu-lang,CHEN Shen School of Information Science and Technology,Northwest University,Xi’an 710127,China

【机构】 西北大学信息科学与技术学院；

【摘要】 本文详细介绍了全变分图像复原的模型,并研究比较了全变分图像去噪的三种数值方法:人工时间演化法,Chan迭代法和固定点迭代算法(fixed point algorithm)。实验表明,全变分模型对被高斯噪声污染的图像有很好的复原结果。三种算法中Chan迭代算法收敛速度最快,计算开销最小,同时能得到最好的复原结果。更多还原

【Abstract】 Image restoration based on total variation model is introduced,and three numerical methods of total variation denoising model are studied and compared in this article:artificial time evolution,Chan algorithm and fixed point algorithm.The experiments show that the total variation model has good restoration result to the image which is polluted by the Gaussian noise.The Chan algorithm converges the quickest,costs the least computation,and gets the best restoration result among three algorithms.更多还原