摘要
超声回波信号中包含大量有关缺陷性质的信息,由于信号在检测和传输过程中不可避免地受到随机噪声干扰,给后续信号处理带来误差。针对该问题,提出改进的希尔伯特-黄变换(Hilbert-Huang Transform,HHT)应用于超声回波信号去噪,提取回波信号中反映缺陷本质特征的信号。改进算法引入小波包变换对超声回波信号作预处理,对处理后的信号进行经验模态分解,提出对数似然检验法,识别并去除噪声引起的虚假分量,重构出有用回波信号。通过对添加高斯白噪声的超声回波信号去噪,验证改进算法应用于超声回波信号去噪的可行性和有效性。
The ultrasonic echo signal contains a large number of information reflects defect properties. Echo signal also has various noise in defecting and transmission that brings error. Improved Hilbert-Huang transform (HHT) is proposed and it is applied to the ultrasonic echo signal de-noising, extraction of echo signal reflects the essential characteristics of the defect. Improved algorithm put forward wavelet packet which doing pretreatment for signal firstly. The log-likelihood test is proposed which was with the log-likelihood value of each intrinsic mode func- tion component and the original signal as the judgment basis to identify and remove illusive components. Then use- ful defect echo signal was reconstructed. Finally, by de-noising the ultrasonic echo signal with Gaussian white noise, the results demonstrate feasibility and effectiveness of the improved HHT.
出处
《科学技术与工程》
北大核心
2013年第12期3481-3487,共7页
Science Technology and Engineering
基金
江苏省"六大人才"高峰高层次人才项目(2010-JXQC-132)
教育部留学回国人员科研启动基金(教外司留[2010]609)
江苏省"青蓝工程"中青年学术带头人资助项目资助
关键词
超声检测
希尔伯特-黄变换
小波包
对数似然检验法
信号去噪
ultrasonic testing Hilbert-Huang transform (HHT) wavelet packets Log-likelihood test(LLT) signal de-noising
作者简介
张永宏(1974-),男,汉族,山东临沂人,博士生导师,教授,研究方向:模式识别与智能系统、图像识别、超声检测等。E—mail:nuist—zyh@yahoo.com.cn。
通信作者简介:郝培培(1987-),女,汉族,江苏徐州人,在读硕士研究生,研究方向:信号处理、相控阵超声无损检测。