| 引用本文: | 张恒,潘仲明.平稳小波去噪算法中的参数选择.[J].国防科技大学学报,2019,41(4):165-170.[点击复制] |
| ZHANG Heng,PAN Zhongming.Parameters selection of stationary wavelet denoising algorithm[J].Journal of National University of Defense Technology,2019,41(4):165-170[点击复制] |
|
| |
|
|
| 本文已被：浏览 6089次 下载 5457次 |
| 平稳小波去噪算法中的参数选择 |
| 张恒1, 潘仲明2 |
|
(1. 国防科技大学 前沿交叉学科学院， 湖南 长沙 410073;2. 国防科技大学 智能科学学院， 湖南 长沙 410073)
|
| 摘要: |
| 为研究平稳小波变换去噪算法在工程应用时如何选择小波基、小波系数处理方法和阈值计算方法的参数以取得最优去噪效果，通过仿真实验，对比不同小波基、不同小波系数处理方法及不同阈值计算方法对平稳小波变换去噪算法去噪效果的影响，对算法的参数选择问题进行研究。实验结果表明：相比其他滤波器组，大部分情况下Daubechies 小波基对应的滤波器组去噪效果更好；信号信噪比较低时选用软阈值法，信噪比较高时选用硬阈值法；使用阈值法处理小波系数，信号信噪比不高的情况下应采取固定阈值法来确定阈值，信号信噪比较高时应采取无偏风险估计法。 |
| 关键词: 平稳小波去噪 多孔算法 参数选择 |
| DOI：10.11887/j.cn.201904023 |
| 投稿日期：2018-03-13 |
| 基金项目:国家自然科学基金资助项目（11804385） |
|
| Parameters selection of stationary wavelet denoising algorithm |
| ZHANG Heng1, PAN Zhongming2 |
|
(1. College of Advanced Interdisciplinary Studies, National University of Defense Technology, Changsha 410073, China;2. College of Intelligence Science and Technology, National University of Defense Technology, Changsha 410073, China)
|
| Abstract: |
| For studying how to select the algorithm parameters, which are wavelet basis, wavelet coefficient processing method and threshold calculation method respectively, in stationary wavelet denoising algorithm to obtain the optimal denoising performance. The denoising performance under different parameters, the selection of orthogonal filter banks and the wavelet coefficients processing method in stationary wavelet denoising algorithm were proposed and compared. The results show that: compared with other filter banks, the filter bank corresponded to Daubechies wavelet basis can achieve a better denoising performance; the soft thresholding method should be selected for processing wavelet coefficient in low signal noise ratio condition, and the hard thresholding method should be utilized when signal noise ratio is high; when using the thresholding method to process the wavelet coefficients, the threshold value should be calculated by the sqtwolog method in low signal noise ratio condition, and rigrsure method should be adopted when the signal noise ratio is high. |
| Keywords: stationary wavelet denoising à trous algorithm parameters selection |
|
|
