文档介绍：一种基于白噪声分解特征的EMD降噪方法
薛志宏1 李广云1 周蓉2
(,河南郑州市陇海中路66号 450052;,江苏南京马群五棵松46号 210049)
摘要经验模态分解算法基于待分解数据本身,避免了小波分解时选取合适小波基函数的困难,具有自适应性。然而传统的经验模态分解降噪是一种强制降噪算法,容易将高频部分的有用信号与噪声一起滤除,从而造成信号失真。针对该问题,在分析白噪声EMD分解特性的基础上,提出了一种EEMD阈值降噪法,利用四组具有不同频谱特征的仿真数据,证明了该算法优于传统的EMD强制降噪法,在消除随机噪声的同时,能够有效保留信号中的高频细节分量,从而缓解了信号的失真。
关键词经验模态分解;白噪声;降噪;阈值法
A Threshold De-noising Method Based On the Characteristics of White Noise
posed by EMD
Xue Zhihong1, Li Guangyun1, Zhou Rong2
(1. Institute of Surveying and Mapping, Information Engineering University, 66 Longhai Road, Zhengzhou, 450052, China
2. 73603 Group,Nanjing, 210049, China)
Abstract: The EMD method is adaptive, with the basis of the position based on and derived from the data, and free from the choice of wavelet base and the determination of the number of position order. An EEMD threshold de-noising method is put forward in this paper to alleviative the drawback of the original EEMD forced de-nosing method which causes distortion in high-ponent. With four type of simulated data with different spectrum, Original EEMD forced de-noising, EEMD threshold de-noising and several kinds of wavelet de-noising methods pared, and the results show that EEMD threshold de-noising method perform better than the forced method due to some significant details within high frequency band are reserved.
Keywords: Empirical Mode position; White Noise