小波变换在脑电特征信号提取中的应用研究

【作者】 王雷；

【导师】 刘伯强；

【作者基本信息】 山东大学 ， 生物医学工程， 2005， 硕士

【摘要】 脑电信号是大脑皮层的神经元具有的生物电活动。脑电信号中包含大量的人体生理与疾病信息,在临床医学上,脑电信号处理不仅可以为某些脑疾病提供诊断的依据,而且还为某些脑疾病提供了有效的治疗手段。正确的提取和识别脑电信号波形,对于疾病的诊断和各种生理现象的解释具有很大的重要性。 由于脑电信号的重要性,自上个世纪20年代出现脑电图记录以来,对于脑电信号的研究就方兴未艾,同时伴随着信号处理技术的发展,对脑电信号的各种处理算法不断涌现,其中傅立叶变换,功率谱估计等方法均应用于脑电信号的处理,这些都极大推进了脑电信号分析的发展。 但是由于脑电信号是非平稳随机信号,所以用傅立叶变换的方法很难得出满意的结论。后来到了二十世纪八十年代,发展起来一门新型的信号处理方法一小波变换理论,小波变换对于信号的高频成分使用逐渐尖锐的时间分辨率以便移近观察信号的快变成分,对于低频成分使用逐渐尖锐的频率分辨率以便移远观察信号的慢变成分。小波的这种既见树木又见森林的信号分析表示特征对分析非平稳信号非常有利。 本文首先介绍了脑电信号的波形的含义和波形提取的导联方法,对各种脑电信号节律作了介绍。主要研究了小波变换在脑电信号提取、消噪、和诱发脑电信号的获取方面的应用,为更加准确的分析脑电信号提供大量时频信息。 对于脑电信号,其中含有大量瞬态信号,这些信号对于疾病的诊断有重要的价值。我们把小波变换应用于这些信号的检测,对于分析瞬态信号和取出噪声取得了明显的效果。采用多分辨率分析的方法对脑电信号进行分解,观测瞬态信号在不同子频带内所表现的时域特性,并将瞬态脉冲较为突出的子带信号组合在一起,构成新的观测信号,然后对新的信号进行检测和定位。用幅值检测法进行去噪,消除噪声干扰,在时频两方面都取得了很好的效果。 同时由于多分辨率分析是在低频段对信号进行细分,而小波包是在低频和高频一起进行细分,这对于检测全频段十分有利,我们用小波包分析方法对脑电信号的主要节律进行了分解,对于进一步仔细研究高频和低频脑电节律的发生机理和疾病波形特征起到了重要作用。 在脑电信号分析中,诱发脑电对于分析脑部疾病和人脑机能有着重要的作更多还原

【Abstract】 Electroencephalogram signal is bioelectric behavior from nerve cell in pallium. much message of health is consisted in the electroencephalogram signal. In clinic, the EGG signals not only offer the evidence of determination of disease, but also offer the effective measurement to cure the diseases. It is very important to get and identity the electroencephalogram signal for determination of disease and explanation of health.Because the electroencephalogram signal is so important, the research of electroencephalogram has been forward from the discovery of electroencephalogram signal in 20s last century. Meanwhile, Arithmetic of electroencephalogram signal process increasingly emerged with the development of technology of signal process, Fourier transformation and power frequency spectrum are applied in electroencephalogram signal process; these means promote the development of electroencephalogram signal process.However electroencephalogram signal is random, it is difficult to get the precious result by Fourier transformation. In 80s twenty century, a new signal process way emerged, which is the wavelet transformation. Little time resolution is adopted to observe fast change of signal for high frequency, and little frequency resolution is adopted to observe slow change of signal for low frequency. Wavelet is the character of watching both tree and forest. This character is advantage to analysis random signal.The content of wave of electroencephalogram signal firstly was introduced in this article and the connected way of getting signal and the difference frequency wave. The application of wavelet transformation in getting electroencephalogram signal were discussed in this article, revulsive visual signal and filter the noise were mainly studied. This produced a lot of message in time and frequency area for analysis electroencephalogram.For electroencephalogram signal, they contain much temporary signal.These signal are very important to detect the disease. We could get the significant result from the detected these signal by means of wavelet transformation. Multi-resolution analysis is adopted in decomposing of electroencephalogram signal, so character of difference frequency range of temporary signals is observed. Pulses of more temporary signals are mixed, and a new signal is consisted by the mixed signal. Then the new signal was detected and positioned. The noise was removed from signal by the means of detecting the extent. The result was obviously availability.Meanwhile, multi-resolution decomposed the signal in low frequency range, and wavelet package decompose the signal in both low and high frequency range; This is the advantage to detect whole frequency range. The electroencephalogram signal to difference frequency range was decomposed by the means of wavelet package. It is very important to research the high and low frequency electroencephalogram signal. It will play a significant role in detecting the disease.The revulsive electroencephalogram signal is important to analysis the disease and function of brain. In this article, we got the revulsive visual electroencephalogram signal hidden in the electroencephalogram, and do the emulator. We got the good result through compared the result.更多还原