基于DSP的电器鸣叫音自动识别系统  

DSP Based automatic Recognition System forhome appliance’s tip sound

作　　者：金乃庆 吴刚 高翠云 Jin Naiqing;Wu Gang;Gao Cuiyun(Whirlpool(China)Global R&d Center,Hefei Anhui,230088;School of Electronics and Information Engineering,Anhui Jianzhu University,Hefei Anhui,230601)

机构地区：[1]惠而浦(中国)全球研发中心,安徽合肥230088 [2]安徽建筑大学电子与信息工程学院,安徽合肥230601

出　　处：《电子测试》2020年第18期34-35,42,共3页Electronic Test

基　　金：安徽省高校省级自然科学研究重大项目(KJ2019ZD56);安徽省自然基金面上项目(180885MF192)。

摘　　要：长期在噪声环境下采用人耳检测蜂鸣器的鸣叫音,易产生听觉疲劳、判断错误,因此,本文设计了基于TI公司的浮点型DSP-TMS320F28335的电器鸣叫音自动识别系统。采用FIR带通滤波器对信号进行滤波,利用FFT提取频谱特征,基于最大相似度进行分类。利用DMA模块设计了乒乓缓存器,保证了实时采集和信号处理同步进行。实验表明:本文的算法可以实现鸣叫音模式的正确识别,且具备FISHER准则的典型特征。It is easy to cause hearing fatigue and mis-judgment for human being to detect the tip sound of the buzzer in the noisy environment for long time.Therefore,We designed an automatic Recognition System for home appliance’s tip sound based on a floating point DSP of TMS320F28335 of Ti in this paper.The FIR bandpass filter was used to filter the signal,using FFT to extract the spectrum features,and using the maximum similarity to classify.Using DMA module todesign the ping-pong cache,and realize the synchronization mechanism of real-time acquisition and signal processing.The experimental results show that the proposed algorithm can recognize the pattern of the sound correctly and meets FISHER’s criterion.

关 键 词：模式识别 TMS320F28335 FFT FIR带通滤波 乒乓缓存 

分 类 号：TN912.3[电子电信—通信与信息系统]