基于Kinect的语音识别技术研究

Speech Recognition Technology of Kinect Sensor

针对Kinect语音识别技术中的语音命令交互控制功能没有得到有效开发的问题,通过对Kinect麦克风阵列的基本原理进行阐述,对音频捕获、音频处理和语音识别过程中主要使用到的应用接口、类、对象、属性、方法和事件等进行深入剖析,以Kinect for Windows Developer Toolkit工具中的Speech Basics-WPF案例中为研究对象,对该应用程序的每一个模块进行单元测试,同时配合静态分析技术,验证了代码的安全性和可靠性,并在此基础上提出了基于Kinect的语音命令实现交互控制的开发流程和编程算法,为Kinect语音命令识别的应用与开发提供策略与方法。

For the issue of the interactive control on Kinect voice commands is not be used effectively,this paper describesthe basic principle of Kinect microphone array,also analyses the application interfaces,classes,objects,properties,methods andevents in the process of audio capture,audio processing and speech recognition.Taking SpeechBasics-WPF case in the Kinect forWindows Developer Toolkit tool as the research object,the paper uses unit tests for each module of the application,at the sametime with the static analysis technology,verifing the security and reliability of the code,also proposes the development process andprogramming algorithm of voice command to achieve interactive control based on Kinect,provides strategy and method for applica?tion and development of command recognition.