基于小波变换的脑电瞬态信号检测

在脑电（ＥＥＧ）信号分析与处理过程中，瞬态信号的检测和定位具有非常重要的实际意义，传统的瞬态脉冲检测方法是匹配滤波，但匹配波滤需要有关瞬态信号的先验知识，因而在实际应用中受一定的限制，本文用小波变换对含有瞬态干扰的脑电信号进行多尺度分解，在某些尺度下，瞬态信号特征得以明显增强，用简单的阈值比较就可以有效地检测并消除瞬态干扰，实验结果表是有，在缺乏先验知识的条件下，小波变换能有效检测出脑电信号中短时，低能量的瞬态脉冲。

Transient Signal Detection ofEEG Based on Wavelet Transform

In the analysis and processing of EEG signal, the detection and the location of transient signal have great practical significance. The traditional approach of transient signal detection is matched filtering. The matched filtering needs the transcendental knowledge of the transient signal, but in many cases, it is impossible to get the transcendental knowledge, so the application of matched filtering can not get ideal results. In this paper, wavelet transform is used to decompose the EEG signal, which contains the interference of transient signals at multi-scales. In some scales of wavelet transform, the characteristics of transient signals are obviously enhanced so that we can accurately locate the positions of transient pulses by using the threshold method, and then remove them. The experimental results show that wavelet transform is an effective approach to detect the short-lived and lower-energy transient pulse under the situation of lacking of transcendental knowledge.