摘要
提出了一种利用非局部相似的图像去噪方法,通过非凸函数阈值得到噪声图像的多个初步去噪图像,其参数的选择较接近,并将这些初步去噪的图像选取相似块,利用非局部平均的方法,对图像进行去噪恢复。通过实验可知,文中提出的方法对于噪声图像去噪具有较好的效果,与典型的非局部平均方法相比,该方法能得到更好的峰值信噪比。
This paper presents a similar use of non-local image denoising methods. Multiple initial denoising images with parameters close to each other are obtained using non-convex function threshold, and then similar images blocks are selected from these preliminary de-noising images for image denoising recovery by non-local averaging method. Experiment shows that the proposed method yields good results with better PSNR than typical non-local averaging methods.
出处
《电子科技》
2014年第2期125-126,共2页
Electronic Science and Technology
关键词
阈值
图像去噪
非局部
threshold value
image denoising
nonlocal
