摘要
图像去噪要求在尽量保留有用信息完整性的基础上去除干扰信息,同时保证去噪结果的时域稳定性。在非局部平均NLM(Non-Local Means)算法的基础上提出了更加合理的图像三维去噪算法。该算法首先进行NLM空域去噪,然后再对空域去噪的结果进行加权平均的时域去噪。实验结果表明,本算法和传统去噪算法相比具有去噪结果稳定并不模糊细节的优点,为实际工程应用提供了技术支撑。
Image denoising removes noise based on keeping useful information as much as possible besides ensuring the stability in time domain of denoising results. This paper presents a more reasonable image three-dimensional denoising algorithm based on the NLM. At First, the NLM is used to remove spatial noise; then, the weighted mean algorithm is used to process the results to remove the noise in time domain. The experimental results show that such algorithm has the advantages of denoising stably without blurring image details compared with the traditional noise reduction algorithms. It also provides a strong technical support for engineering application.
出处
《红外技术》
CSCD
北大核心
2013年第4期238-241,共4页
Infrared Technology
关键词
非局部平均NLM
结构相似度
图像三维去噪
non-Local Means, structure similarity, image three-dimensional denosing