城市噪声监测平台的设计与实现

城市噪声在不同程度上影响人们的日常生活。为了全面监测城市中的噪声,分析其时空分布特征及其类别,以便针对性实施治理,研发了一种基于公共移动载体上低成本综合监测终端与监测系统。城市噪声监测终端采用低成本开发板Sed Board为处理单元,采用i436噪声传感器,该开发板上有板载的主控芯片Zynq-7000 ALL Programmable SoC集成ARMCortex-A9双核以及最多可达500万多个逻辑门的可编程逻辑单元能够灵活地用于各种目标应用,采用其ALSA音频架构实现噪声数据采集,利用ZedBoard自带网口将音频数据传输到pc端,在pc端通过基于PyTorch框架的声谱分析算法,将采集的噪声识别为载体噪声、交通噪声、商业噪声、生活噪声3类主要噪声,测试集上分类准确率为98.2700%。

Mobile monitoring platform for urban noise design and implantation

Urban noise affects people′s daily life to varying degrees.In order to comprehensively monitor the noise in the city,analyze its spatial and temporal distribution characteristics and categories,in order to implement targeted governance,a low-cost integrated monitoring terminal and monitoring system based on public mobile carriers was developed.Sed Board,a low-cost development board,is used as the processing unit and i436 noise sensor is used in the urban noise monitoring terminal.The main control chip Zynq-7000 ALL Programmable SoC on the development board,which integrates ARM Cortex-A9 dual-core and programmable logic units with up to 5 million logic gates,can be flexibly used in various target applications.The noise data acquisition is realized by using its ALSA audio architecture,and the audio data is transmitted to the PC terminal through ZedBoard′s own network port.At the PC terminal,the collected noise is recognized as carrier noise,traffic noise,commercial noise and living noise by using the spectrum analysis algorithm based on PyTorch framework.The classification accuracy of the test set is tested.98.2700%.