针对噪声环境下麦克风系统识别率降低的问题,提出一种基于双传感器特征融合的语音识别系统。利用STM32单片机同时采集说话人发声时的皮肤振动语音信息和麦克风语音信息,通过WiFi发送至上位机,将双路语音特征MFCC参数融合并与隐马尔可夫模型结合用于孤立词识别研究。实验结果表明,在安静和噪声环境下,与单一麦克风语音识别系统相比,此系统具有更高的识别率和更强的抗噪能力,鲁棒性更好。
Abstract
Aiming at the problem that the recognition rate of microphone system decreases in noisy environment,a speech recognition system based on dual-sensor feature fusion is proposed.The STM32 single-chip microcomputer is used to simultaneously collect the skin vibration voice information and the microphone voice information of the speaker and send them to the host computer through WiFi.The MFCC parameters of the two-way voice feature are fused and combined with the hidden Markov model for isolated word recognition research.The experiment results show that compared with the single microphone speech recognition system,this system has higher recognition rate,stronger anti-noise ability and better robustness in quiet and noisy environments.
关键词
STM32 /
麦克风语音 /
隐马尔可夫模型 /
孤立词识别 /
ESP8266
Key words
STM32 /
microphone speech /
hidden Markov model /
isolated word recognition /
ESP8266
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 黄志东.鲁棒性语音识别技术研究综述[J].信息通信,2019(11):20-22.
[2] Graciarena M,Franco H,Sonmez K,et al.Combining standard and throat microphones for robust speech recognition[J].IEEE Signal Processing Letters,2003,10(3):72-74.
[3] Radha N,Shahina A,Nayeemulla Khan A.Improving Recognition of Speech System Using Multimodal Approach[C]//International Conference on Innovative Computing and Communications. Springer, Singapore,2019:397-410.
[4] 张海宁.强噪声环境下语音识别方法的研究[D].哈尔滨:黑龙江大学,2021.
[5] 吴卫鹏.基于改进谱减的语音增强算法研究[D].南京:南京邮电大学,2019.
[6] 陈锡锻.一种双门限语音端点检测算法[J].浙江工贸职业技术学院学报,2021,21(2):43-46.
[7] 胡洋霞.基于DTW模型的非特定人孤立词语音识别研究[D].天津:河北工业大学,2015.
[8] 曹冠彬.基于HMM的连续语音识别技术研究[D].南京:南京理工大学,2018.
[9] Cheng X,Duan Q.Speech emotion recognition using gaussian mixture model[C]//The 2nd international conference on computer application and system modeling,2012:1222-1225.
基金
* 气体钻井安全监测的前兆预警关键传感器研究(61731016)。
PDF(1638 KB)
