摘要
为了在消除图像处理中的噪声的同时尽可能保留图像细节信息,研究了带钢表面图像的迭代控制核回归图像去噪模型;在以欧氏距离决定权值的基础上,考虑像素灰度值,引入一个迭代过程,并对全局平滑参数进行有效的选取,通过主观评价和客观评价对该方法在带钢表面图像的去噪效果做出评价。实验结果表明,基于迭代控制核回归的图像去噪方法相较于传统的图像去噪方法提高了图像的信噪比,在有效去除噪声的同时更好的保存了图像的纹理和边缘等细节信息。
In order to preserve the details of image possibly while eliminating noise of image processing, the image de-noising model of strip surface based on iterative steering kernel regression was studied, which is based on the Euclidean distance.It pulls into an iretative process by considering the gray value and selects the global smoothing parameter. The effection of the strip steel surface de-noising was evaluated by subjective evaluation and objective evaluation~ The experimental results show that, compared with the traditional image denoising method ,the image de-noising method based on lterative Steering Kernel Regression improves the SNR of the image. The image texture and edge detals were better preserved while the noise was effectively removed.
出处
《机械设计与制造》
北大核心
2017年第5期119-122,共4页
Machinery Design & Manufacture
基金
冶金装备及控制教育部重点实验室基金(2013B01)
湖北省教育厅基金(B2013242)
武汉市科技攻关项目资助(2014010202010088)
关键词
图像去噪
带钢
表面缺陷
迭代控制核回归
Image Denoising, Strip Steel, Surface Defect,Iterative Steering Kernel Regression
