阵列麦克风声源定位技术在师范生授课过程中的应用研究
|
摘要
提高师范生的授课技能是师范院校培养计划中的重要内容,师范生的授课主要采取两种形式实现老师与学生的互动,一是肢体动作,二是语言语音。利用语音采集设备记录课堂的语音信息,从语音信息中提取多种参数,可以反映出师范生授课情景的整体氛围,为提高师范生上课过程中的交互性起到了重要的作用。论文利用微软公司开发的KINECT设备,利用该设备的阵列麦克风音源定位技术实现声音属性的记录,存储与处理。对语音的方向,强弱,还有整个课堂的底噪的判断达到了很好的效果。论文利用Kinect摄像头的4个麦克风所构成的麦克风阵列,采用传感器数字信号处理器,麦克风阵列分别对某个发音源所接收到的音量强弱,加上噪声抑制和消除回音功能,在VisualStudio2015环境搭载KinectforWindows SDKV2.0平台驱动Kinect2.0,以此分析出水平方向的声音来源以及该声源的特征。利用这个功能,我们应用于师范生授课过程中分析师范生授课声音的相关参数,以此来通过采集到师范生授课过程中声音信息的相关数据做出一个评判。
Improving the teaching skills of normal students is an important part of the training plan of normal colleges. The teaching of normal students mainly adopts two forms to realize the interaction between teachers and students. One is physical movement and the other is language speech. Using the voice acquisition device to record the voice information of the classroom, and extracting various parameters from the voice information, can reflect the overall atmosphere of the teacher's teaching situation, and play an important role in improving the interactivity of the normal students in the course of class. The paper uses the KINECT device developed by Microsoft Corporation to realize the recording, storage and processing of sound attributes by using the array microphone sound source localization technology of the device. The direction of the voice, the strength and weakness, as well as the judgment of the noise level of the entire classroom have achieved good results. The paper uses the microphone array of the four microphones of the Kinect camera, and uses the sensor digital signal processor to filter the background noise and enhance the clarity when the sound is received. The volume of the microphone array is received by a certain pronunciation source. On the noise suppression and echo cancellation function, the Kinect for Windows SDK V2.0 platform is used to drive Kinect2.0 in the Visual Studio 2015 environment to analyze the horizontal sound source and the characteristics of the sound source. Using this function, we apply the relevant parameters of the voices of the analysts in the course of the teacher's teaching, so as to make a judgment by collecting the relevant data of the sound information during the teaching process.
Improving the teaching skills of normal students is an important part of the training plan of normal colleges. The teaching of normal students mainly adopts two forms to realize the interaction between teachers and students. One is physical movement and the other is language speech. Using the voice acquisition device to record the voice information of the classroom, and extracting various parameters from the voice information, can reflect the overall atmosphere of the teacher's teaching situation, and play an important role in improving the interactivity of the normal students in the course of class. The paper uses the KINECT device developed by Microsoft Corporation to realize the recording, storage and processing of sound attributes by using the array microphone sound source localization technology of the device. The direction of the voice, the strength and weakness, as well as the judgment of the noise level of the entire classroom have achieved good results. The paper uses the microphone array of the four microphones of the Kinect camera, and uses the sensor digital signal processor to filter the background noise and enhance the clarity when the sound is received. The volume of the microphone array is received by a certain pronunciation source. On the noise suppression and echo cancellation function, the Kinect for Windows SDK V2.0 platform is used to drive Kinect2.0 in the Visual Studio 2015 environment to analyze the horizontal sound source and the characteristics of the sound source. Using this function, we apply the relevant parameters of the voices of the analysts in the course of the teacher's teaching, so as to make a judgment by collecting the relevant data of the sound information during the teaching process.
引文
[1]韩雪,李泽滔,孙昊.基于Kinect传感器的移动机器人声源目标跟踪系统[J].自动化与仪器仪表,2015(06):185-186.
[2]赵圣,崔牧凡,尤磊,王鸿鹏.基于小型麦克风阵列的声源定位技术[J].华中科技大学学报(自然科学版),2013,41(S1):188-191.
[3]张自嘉,李贺,花晓蕾,张丽萍.基于波束形成算法的声源定位与DSP实现[J].压电与声光,2015,37(01):117-121.
[4]朱虹.基于Kinect的特定说话人跟踪系统的研究与实现[D].延边大学,2017.
[5]王书海,石俊峰.基于麦克风阵列的声源定向时延估计的研究[J].科技创新与应用,2018(23):13-14.
[6]王森.Kinect体感程序设计入门:使用C#和C++.北京:科学出版社,2014.11
[2]赵圣,崔牧凡,尤磊,王鸿鹏.基于小型麦克风阵列的声源定位技术[J].华中科技大学学报(自然科学版),2013,41(S1):188-191.
[3]张自嘉,李贺,花晓蕾,张丽萍.基于波束形成算法的声源定位与DSP实现[J].压电与声光,2015,37(01):117-121.
[4]朱虹.基于Kinect的特定说话人跟踪系统的研究与实现[D].延边大学,2017.
[5]王书海,石俊峰.基于麦克风阵列的声源定向时延估计的研究[J].科技创新与应用,2018(23):13-14.
[6]王森.Kinect体感程序设计入门:使用C#和C++.北京:科学出版社,2014.11