Gaussian Noise Removal Method Based on Empirical Wavelet Transform and Hypothesis Testing
2022 3rd International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering (ICBAIE)(2022)
摘要
Most industrial signals in reality have nonlinear and non-stationary characteristics, so the effect of signal denoising by general signal analysis methods is often not ideal. Taking the vibration signal of rolling bearing as an example, statistically, it is non Gaussian distribution, and the mixed noise is mostly Gaussian noise, and the noise bandwidth is wide, so the traditional filtering method is not suitable. Based on the above analysis, a Gaussian noise elimination method based on empirical wavelet transform and hypothesis test is proposed in this paper .Taking the vibration signal of rolling bearing as an example, Gaussian noise is mixed into it. The signal is decomposed into several components through empirical wavelet transform, and which components meet the Gaussian distribution are identified as noise through the hypothesis test of Gaussian distribution. The simulation results show that the denoising method proposed in this paper is more stable than the traditional method.
更多查看译文
关键词
empirical wavelet transform,hypothesis test,Gaussian noise
AI 理解论文
溯源树
样例
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要