何睿　胡定玉　于淼　李晓龙

摘 要： 声源可视化技术对传声器数量和采集系统的性能具有一定的要求，传统驻极体传声器阵列采集系统存在功耗高、体积大、采集系统可移植性差和拓展性差等问题。本文设计了36阵元数字MEMS传声器阵列成像系统，有效克服了以上问题。采集模块选用FPGA为硬件平台，完成数据采集和信号处理模块设计，分析了I2S音频传输协议的实现、数据传输机制的建立和上位机数据的交互方式。并在普通室内环境中对此系统进行实验验证，系统的声源成像效果良好、成像分辨率高，在实际的噪声源定位与噪声设备监测中有广泛的应用前景。

关键词： 声源可视化; 波束形成; FPGA; MEMS声传感器

文章编号： 2095-2163（2021）07-0222-06中图分类号：TP212，TB52+5文献标志码： A

Design of microphone array imaging system based on FPGA

HE Rui1，  HU Dingyu1，2， YU Miao3， LI Xiaolong3

（1 School of Urban Railway Transportation， Shanghai University of Engineering Science， Shanghai 201620， China; 2  Shanghai

Engineering Research Center of Vibration and Noise Control Technologies for Rail Transit， Shanghai University of Engineering

Science， Shanghai 201620， China; 3 China Railway Harbin Bureau Group Co.， Ltd. Harbin EMU， Harbin 150000， China）

【Abstract】Sound source visualization techniques have some requirements on the number of microphones and the performance of the data acquisition system. Traditional electret microphone array data acquisition systems have problems such as high-power consumption， large size， poor portability and poor expandability of the data acquisition system. In this paper， a 36-array digital MEMS microphone array imaging system is designed to effectively overcome the above problems. The FPGA is selected as the hardware platform for the data acquisition module， and the data acquisition and signal processing modules are designed. The implementation of I2S audio transmission protocol， the establishment of data transmission mechanism and the interaction mode of the upper computer data are analyzed. The system has been experimentally verified in a common environment， and the system has good sound source imaging effect and high imaging resolution， which has a wide application prospect in practical noise source localization and noise equipment monitoring.

【Key words】sound source visualization; beamforming; FPGA; MEMS microphone

0 引 言

聲源成像可视化技术[1]利用传声器阵列采集多通道声压数据，通过信号处理算法，将声信号转换成图像信号使得空间声源分布可视化，可快速定位声源位置、分布情况以及瞬时动态变化特征，以用于设备状态评估和监测。目前，传声器阵列采集系统可分为2种类型。一种是由数据采集卡、标准总线和计算机构成，主要用于工业测试环境。其中，以美国国家仪器（NI）、泰克（Tektronix）、安捷伦（Agilent）为首的公司生产的仪器采集性能优越，工作状态稳定，但该类型采集仪器具有高功耗、大体积的缺点[2]，且需要配合相应的数据采集卡使用，具有专一性强、可移植性差等局限性。另一种是由微处理器为采集核心的嵌入式采集系统构成，主要用于实验测试和部分工业测试。……