一种基于图像边缘检测的全变分的去噪方法  

Total Variation Image Denoising Based on Image Edge Detection

作　　者：黄伟[1] 王书文[2] 

机构地区：[1]西北民族大学中国民族信息技术研究院,甘肃兰州730030 [2]西北民族大学电气工程学院,甘肃兰州730030

出　　处：《计算机技术与发展》2009年第2期24-26,30,共4页Computer Technology and Development

基　　金：国家自然科学基金项目(60675059)

摘　　要：提出了一种基于边缘检测的全变分图像去噪方法。在利用全变分去噪之前,先用Canny算子检测图像的边缘,对检测出的边缘区域和非边缘区域做标记;然后在边缘和非边缘区域设置不同的均衡系数,利用全变分模型对图像进行去噪。实验结果表明该算法能抑制以往全变分模型方法产生的阶梯效应,具有较好的图像恢复效果。Presents a total variation（TV） image denoising based on edge detection. Before denoising in the use of TV model, the edge of a noised image was detected by Canny operator, and marked the detected image edge region and non- edge region. Then the different proportion coefficients were set in image edge and non - edge region, using TV model to carry on image denoising. Experiments show that the proposed method can not only restrain staircasing effect occurred in the previous TV model, but also has great restored effect.

关 键 词：图像去噪 CANNY算子 全变分 边缘检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]