改进小波降噪算法在轴承缺陷图像的应用

发布时间：2018-03-01 18:39

本文关键词： 小波分解维纳滤波阈值估计图像降噪峰值信噪比　出处：《组合机床与自动化加工技术》2017年11期 　论文类型：期刊论文

【摘要】：针对传统Bayes阈值不能随小波分解尺度变换以及提高传统算法图像降噪效果的问题,文章提出一种改进的基于小波维纳滤波与Bayes自适应阈值估计图像降噪算法,该算法在多层小波变换的基础上,对小波分解后的第一层细节系数进行维纳滤波处理,对其他层细节系数进行改进Bayes软阈值估计算法处理,最后对处理后的小波系数进行重构,得到降噪图像。实验结果表明,该方法在图像峰值信噪比(PSNR)定量指标上优于传统的小波Bayes软阈值估计图像降噪方法,并将该方法成功的应用于轴承缺陷图像的降噪预处理以及轴承缺陷图像边缘检测中,达到了图像降噪的优化效果。
[Abstract]:Aiming at the problem that the traditional Bayes threshold can not be transformed with the wavelet decomposition scale transform and improves the image denoising effect of the traditional algorithm, an improved image denoising algorithm based on wavelet Wiener filter and Bayes adaptive threshold estimation is proposed in this paper. On the basis of multi-layer wavelet transform, the first detail coefficient of wavelet decomposition is processed by Wiener filter, the other layer detail coefficients are processed by improved Bayes soft threshold estimation algorithm, and the wavelet coefficients are reconstructed. The experimental results show that the proposed method is superior to the traditional wavelet Bayes soft threshold estimation method in image denoising. The method is successfully applied to the pre-processing of bearing defect image and the edge detection of bearing defect image, and the optimal effect of image de-noising is achieved.
【作者单位】：北京信息科技大学现代测控技术教育部重点实验室;
【基金】：国家高技术研究发展计划(863计划)(2015AA043702) 北京市教委科技计划重点项目(KZ201611232032);北京市教委科研计划项目(KM201611232020)
【分类号】：TH133.3;TP391.41

【相似文献】