基于NLMS自适应滤波的实时声音信号检测识别

摘要

在特定区域环境中，提出了一种基于NLMS自适应滤波的实时声音信号检测识别方法。该方法NLMS对滤波后的误差信号进行双门限检测，并以通过检测的疑似信号所持续的时间，作为其是否为目标信号的评判指标。通过对待处理声音文件的测试，该方法可以有效地去除非期望信号，保留目标信号，从而实现对目标信号的准确检测。%In a particular region environment, proposes a real-time voice signal detection and recognition method based on NLMS adaptive filter. The method of double threshold detection of the error signal after NLMS filtering, and through the suspected signal detection for extended periods of time, as if it is the evaluation index of the target signal. By treating the sound file processing test, the method can effectively remove the nonexpected signal, preserve the target signal, in order to achieve accurate detection of target signal.