【摘要】 瞬态信号的检测在信号处理研究中占有重要地位。本文在多尺度分解的基础上对含噪声信号进行尺度上和尺度间作两维自适应滤波，并采用逐级消除平均误差的方法对自适应滤波算法进行了改进，从而有效地滤去噪声，达到提取信号的目的。基于瞬态信号的特点，在滤波的方法上提出了反向滤波的方法。实践结果表明，该方法在信噪比较小时仍能取得较好的结果。更多还原

【Abstract】 Detection of transient signal plays an important role in signal processing. The temporal signal with noise is adaptively filtered in two dimensions with multi scale analysis. It improves the algorithm with the method calculating mean error so that the noise is effectively derogated. According to the character of the transient signal, a new anti filtering method is prompted so that the loss of the orginal information because of the weight system′s non steadiness is reduced. The result shows that this method is effective even though the signal noise ratio is low.更多还原