基于边缘提取的非局部均值图像去噪

Non-local mean image denoising based on edge extraction

图像去噪是图像处理和计算机视觉中的一个研究热点,其目的是从噪声图像中还原原始图像,且尽可能地保留更多的图像细节。非局部均值去噪算法（NLM）是近年来去噪效果比较出色的算法之一,但是传统的NLM算法在对边缘和纹理较丰富的图像进行去噪时,容易过平滑从而丢失很多细节信息。针对此问题,提出了一种基于边缘提取的NLM算法。该方法通过边缘提取,将图像边缘部分与平坦部分像素分开计算,减少不相似图像块之间的干扰,使得结构相似的邻域获得更大的权值,从而提高去噪效果的同时,保留更多的图像细节。实验结果表明：本文算法的去噪性能在主观评价和客观评价准则下都取得了较好的效果,去噪后的图像效果良好,且保留更多的细节。

Among image denoising is a hotspot in image processing and computer vision,with the aim of restoring the original image from the noise image and preserving as much detail as possible.Non-local mean denoising algorithm is one of the best algorithms for denoising in recent years.However,the traditional NLM algorithm is easy to smoothen the image when it is denoised at the edge and texture.Aiming at this problem,this paper proposes an NLM algorithm based on edge extraction.In this method,the edge of the image is separated from the flat part by the edge extraction,and the interference between the image blocks is reduced,so that the neighborhood of the similar structure can obtain more weight,so as to improve the denoising effect.More detail of the image.The experimental results show that the denoising performance of this algorithm has achieved better results under the subjective evaluation and objective evaluation criteria.The image after the denoising is good and retains more details.