CAJ | 학술논문

　　低压配电网的信道环境极为恶劣，语音信号在信道内传输受诸多干扰因素的影响。采用小波变换去除语音信号中的脉冲噪声效果不理想，导致信号传输的可靠性差。脉冲噪声特点是随着脉冲噪声尺度减小其模值迅速增加而语音信号正好相反，因此利用小波模极大值算法选取合理的尺度就可以把语音信号中的脉冲噪声去除掉。基于此，本文中选用小波变换与小波模极大值相结合的方法对含噪声的语音信号进行去噪处理。仿真结果表明，这种结合方法对含有噪声的语音信号去除脉冲噪声效果显著，相比较小波硬阈值算法输出信噪比也有所提高。
　　저압배전망적신도배경겁위악렬，어음신호재신도내전수수제다간우인소적영향。채용소파변환거제어음신호중적맥충조성효과불이상，도치신호전수적가고성차。맥충조성특점시수착맥충조성척도감소기모치신속증가이어음신호정호상반，인차이용소파모겁대치산법선취합리적척도취가이파어음신호중적맥충조성거제도。기우차，본문중선용소파변환여소파모겁대치상결합적방법대함조성적어음신호진행거조처리。방진결과표명，저충결합방법대함유조성적어음신호거제맥충조성효과현저，상비교소파경역치산법수출신조비야유소제고。
The speech signal is affected by various factors for harsh transmission condition over low voltage distribution network channel. The wavelet method for removing pulse noise is not so effective that the signal has low reliability. The pulse noise, which can be effectively removed from the speech signal by applying wavelet modulus maximum algorithm, has the characteristic of the modulus increase and noise reduction. But the speech signal is just the opposite. This paper proposes that the algorithm combining wavelet transform with wavelet modulus maximum disposes the speech signal in the low voltage distribution network channel. The simulation results show that this method is remarkably effective.