摘要
针对传统智能家电离线语音识别方法识别效率低、识别准确率差的问题,提出了一种在文本信息分词处理下的智能家电离线语音识别方法,对获取的语音信号进行傅里叶变换,得到语音的频谱,通过回声消除和降噪去除冗余信息,在远场语音识别中,智能家电往往离用户的距离较远,所获取的语音信号常常在远距离的传播过程中发生衰减,因此,对信号进行增强处理,将命令词和自然语言转换成两种文本形式,实现智能识别。为验证识别方法效果,进行对比实验,结果表明,文本信息分词处理下的智能家电离线语音识别方法具有很强的识别能力,识别准确率接近于99%,能够快速地识别用户对智能家电的指令,而且不需要全网覆盖,在离线的情况下也能正常工作。
Aiming at the problems of low recognition efficiency and poor recognition accuracy of traditional off-line speech recognition methods for intelligent appliances,an off-line speech recognition method for intelligent appliances is proposed based on word segmentation of text information.The obtained speech signals are Fourier transformed to obtain the speech spectrum,and redundant information is removed by echo cancellation and noise reduction The distance from the user is far away,and the acquired voice signal often decays in the process of long-distance transmission.Therefore,the signal is enhanced,and the command word and natural language are converted into two text forms to realize intelligent recognition.In order to verify the effect of the recognition method,comparative experiments are carried out.The results show that the intelligent home appliances offline speech recognition method based on text information segmentation processing has a strong recognition ability,and the recognition accuracy is close to 99%.It can quickly recognize the user’s instructions to the intelligent home appliances,and it does not need the whole network coverage,and can work normally under the offline condition.
作者
宋雪亚
王传安
SONG Xueya;WANG Chuanan(Anhui Science and Technology University,Bengbu Anhui 233000,China)
出处
《自动化与仪器仪表》
2020年第12期161-164,共4页
Automation & Instrumentation
基金
安徽省教育厅自然科学重点项目:超密集网络中基于协同基站簇的无缝切换机制研究(No.KJ2016A176)
安徽科技学院科研项目:互联网+技术下的无线移动视频智能监控技术研究(No.ZRC2016495No.)
安徽省优秀拔尖人才培育项目,高校优秀青年人才支持重点项目(No.gxyqZD2018069)。
关键词
文本信息分词
分词处理
智能家电
离线语音
识别方法
text information word segmentation
word segmentation processing
smart appliances
offline speech
recognition method
