基于麦克阵列的点声源定位及声源类型分析

发布时间：2018-08-07 13:20

【摘要】：基于麦克阵列的点声源定位是近年来声源定位的研究热点,在军事领域和语音增强领域应用广泛,但在环境噪声监测方面的应用相对较少。目前主要的麦克风阵列模型有平面式麦克阵列和立体式麦克阵列两种。平面式阵列按空间摆放位置形式可分为:一字型、T型、圆型等。立体式阵列可多方考虑声源的方位,在判断角度上可以更加精确,然而三维的快速傅里叶变换及相关算法,增加了运算上的复杂程度和花费的时间。平面式阵列一定程度上减少了计算量,但在检测高度方面受到了相应的限制。本文在研究前人声源定位方法基础上,提出了一种基于半球型麦克风阵列的声源定位系统。首先归纳总结了现有声源定位算法的优缺点,结合三维复合声场的分布规律,针对目前麦克风阵列定位算法计算量大和不便于移植的特点,提出并设计了半球型声源定位方案,给出了设计的理论依据和定位算法推演;其次,采用STM32嵌入式处理器对麦克风阵列数据进行采集和处理,包括声音信号的归一化处理和通过FFT变换得到声音信号的频率信息,并将频率引入噪声声源分析,通过无线传输模块将处理后的数据上传至上位机做进一步定位处理,并设计了数据通讯编码,实现数据的高效传输;最后,对设计完成的系统进行了定位实验,分别对不同位置下的声源进行了定位测试,给出了测试数据,采用LabVIEW虚拟仪器完成了声源的定位,对半球型声源定位系统的误差进行了分析,得到了良好的定位效果。本文给出的声源定位系统噪声测量分辨率为0.1dB,系统定位误差为5%,频率测量范围为100Hz~10kHz。系统良好的实现了点声源的三维定位,并对声源的类型区分做出了初步研究,为声源定位的后续研究提供了新的解决思路。
[Abstract]:Point sound source location based on Mike array is a hot topic in recent years. It is widely used in military field and speech enhancement field, but it is seldom used in environmental noise monitoring. At present, there are two kinds of microphone array models: planar microphone array and stereoscopic microphone array. The plane array can be divided into one font T type, circle type and so on according to the space position form. The stereo array can consider the direction of the sound source in many ways, so it can be more accurate in judging angle. However, the 3D fast Fourier transform and related algorithms increase the complexity and time of the calculation. The planar array reduces the computational complexity to some extent, but it is limited in height detection. In this paper, a sound source location system based on hemispherical microphone array is proposed based on the previous methods of sound source location. Firstly, the advantages and disadvantages of the existing sound source localization algorithms are summarized. According to the characteristics of the current microphone array localization algorithm, which is large in computation and is not easy to transplant, a hemispherical sound source location scheme is proposed and designed, combined with the distribution law of 3D composite sound field. The theoretical basis of the design and the derivation of the location algorithm are given. Secondly, the microphone array data is collected and processed by STM32 embedded processor, including the normalized processing of the sound signal and the frequency information of the sound signal obtained by the FFT transform. The frequency is introduced into the noise source analysis, and the processed data is uploaded to the upper computer through the wireless transmission module for further positioning processing, and the data communication code is designed to realize the efficient data transmission. The localization experiment of the designed system is carried out, the sound sources in different positions are tested, the test data are given, and the sound source location is accomplished by using LabVIEW virtual instrument. The error of the hemispherical sound source positioning system is analyzed and a good localization effect is obtained. The noise measurement resolution of the sound source positioning system is 0.1 dB, the system positioning error is 5, and the frequency measurement range is 100 Hz ~ 10 kHz. The system realizes the 3D location of the point sound source well, and makes a preliminary study on the classification of the sound source, which provides a new solution for the further research of the sound source location.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN912.3

【相似文献】