噪声下基于耳蜗基底膜分频特性的声源定位

发布时间：2018-08-28 16:19

【摘要】：在人与人交流中,人耳能很好地定位出周围声音的位置,而在声源定位技术里,基于传统麦克风声源定位在噪声环境下无法精确地定位出声源。根据人耳能准确地辨别出声源这一特性,提出一种在噪声下基于人耳耳蜗基底膜分频特性的声源定位方法。该方法是利用多个麦克风对声源信号进行采集,将采集到的信号通过基底膜滤波器进行滤波、除去与声源无关的噪声,运用空间映射互相关方法对声源进行定位。该方法所定位出的声源位置比传统声源定位技术所定位的位置误差小、精准度高。实验结果表明,噪声环境下基于耳蜗基底膜分频特性的声源定位比基于传统麦克风声源定位技术更接近声源的真实位置;噪声环境下基于耳蜗基底膜分频特性的声源定位具有更高的精度与更好的鲁棒性。
[Abstract]:In the communication between people, the human ear can locate the sound position of the surrounding sound very well, but in the sound source localization technology, the sound source can not be accurately located in the noise environment based on the traditional microphone sound source localization. According to the characteristic that the human ear can accurately distinguish the sound source, a sound source location method based on the frequency division characteristic of the basal membrane of the cochlea under noise is proposed. This method uses several microphones to collect the sound source signal, filters the collected signal through the basement membrane filter, removes the noise independent of the sound source, and uses the spatial mapping cross-correlation method to locate the sound source. The position of sound source located by this method is smaller and more accurate than that of traditional sound source location technology. The experimental results show that the sound source location based on cochlear basement membrane frequency division is closer to the real location of the sound source than that based on the traditional microphone sound source location technology. The acoustic source location based on cochlear basement membrane frequency division in noisy environment has higher accuracy and better robustness.
【作者单位】：重庆邮电大学自动化学院;重庆邮电大学光电工程学院;
【基金】：国际合作项目(2010DFA12160)~~
【分类号】：TN912.3

【参考文献】