论文部分内容阅读
为了消除波束形成声源识别时地面反射对声源识别结果的影响,给出了基于除自谱的互谱成像函数的镜像地波束形成方法.以半圆形阵列为例,基于传统波束形成方法和镜像地波束形成方法模拟的结果表明,低频时镜像地波束形成声源主瓣近似呈圆形,更接近真实声源形状,在中高频处镜像地波束形成方法相对于传统波束形成方法旁瓣水平被有效衰减,声源定位精度更高.进一步,研究了传统波束形成方法和镜像地波束形成方法在地面不同吸声特性情况下的声源识别性能,结果表明,当吸声系数较低时,镜像地波束形成仍能有效识别声源,当地面的吸声系数较高时,镜像地波束形成方法的声源识别性能变差.最后基于扬声器声源的算例试验验证了上述模拟仿真结论的正确性及镜像地波束形成在识别实际声源时的有效性.
In order to eliminate the influence of ground reflection on the sound source recognition results when beamforming sound source recognition, a mirror image beamforming method based on the cross-spectral imaging function in addition to the self-spectrum is presented.With the semi-circular array as an example, based on the traditional beamforming method The results show that the main lobe of the sound source in the mirror image is approximately circular, which is closer to the shape of the real sound source at low frequency. Compared with the traditional beamforming method, The level is effectively attenuated and the localization accuracy of the sound source is higher.Furthermore, the recognition performance of the sound source under different sound absorption characteristics of the ground by the traditional beamforming method and the image-based beamforming method is studied. The results show that when the sound absorption coefficient is low , The beamforming of the mirror can still identify the sound source effectively, and when the sound absorption coefficient of the ground is high, the sound source recognition performance of the mirror beamforming method becomes worse.At last, the example experiment of speaker sound source verifies the above simulation results The correctness and the validity of the imaged beamforming in recognizing the actual sound source.