面向可穿戴计算机音频的环境感知系统的研究

发布时间：2018-02-21 08:48

本文关键词： 可穿戴计算机声音识别环境感知 HMM　出处：《哈尔滨工业大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着计算机性能的不断更新升级，个人计算机已经逐渐向小型化微型化发展，可穿戴计算机就是微型计算机的一个重要代表可穿戴计算目前已经成为了各大互联网企业研究的热点，各自都推出了相关的可穿戴计算产品，比较典型的有Google的智能眼镜ξProject Glassο，三星的智能手表ξGalaxy Gearο，这些产品目前都以上市另外在一些专业领域也有许多可穿戴计算的产品，特别是在医疗和军事领域声音作为自然界传播信息的重要媒介，，它给人们带来大量的关于环境的信息声音识别技术在的可穿戴计算机上的应用，符合可穿戴计算机发展的方向，即突出增强人的能力及辅助人的生产活动声音识别技术的应用增强了人的听觉能力，同时在分析音频信号时反馈出发声物体的信息在有联网功能的可穿戴计算机上，每台可穿戴计算机可作为网络中的节点，将采集的信息通过网络中心同其他节点进行共享，并由中心整合分析大范围环境的信息，为可穿戴计算机增添环境感知的功能本文查阅了国内外相关研究的大量文献，对可穿戴计算机及声音识别领域的发展和现状进行了解，分析了声音信号与语音信号的不同，确定了开发声音识别系统的技术基础和框架要求通过研究声音识别理论，提出了声音信号的预处理特征提取和基于HMM模型的识别技术，并根据可穿戴计算机的特点，结合环境感知的目的，确定适用的声音识别过程整个系统的框架结构及各模块的功能和设计本系统采用跨平台的C++图形用户界面应用程序Qt实现各模块功能的开发，完成系统的整合最后测试了系统的实际使用情况，对系统的性能进行分析，总结了系统的可取之处以及还需要改进的地方
[Abstract]:With the continuous updating and upgrading of computer performance, personal computers have gradually moved towards miniaturization. Wearable computer is an important representative of microcomputer wearable computing has become a hot spot in the research of various Internet enterprises, each of which has introduced related wearable computing products. Typical of them are Google's smart glasses 尉 Project Glass o, Samsung's smartwatch 尉 Galaxy Gear, which are now on the market. In addition, there are many wearable computing products in some professional fields, especially in the medical and military fields. Sound, as an important medium for the dissemination of information in nature, has brought people a large number of applications of environmental information sound recognition technology in wearable computers, which is in line with the development direction of wearable computers. That is to say, the application of the technology of enhancing human's ability and assisting human's production activity sound recognition enhances human's auditory ability, and when analyzing audio signal, the information of audible object is feedback on wearable computer with network function. Each wearable computer can be used as the node in the network. The collected information can be shared with other nodes through the network center, and the center will integrate and analyze the information in a wide range of environments, so as to add the environment-aware function to the wearable computer. In this paper, a large number of domestic and foreign literatures are reviewed, and the development and present situation of wearable computer and sound recognition field are discussed, and the difference between sound signal and speech signal is analyzed. The technical foundation and framework of developing sound recognition system are determined. By studying the theory of sound recognition, the preprocessing feature extraction of sound signal and the recognition technology based on HMM model are put forward, and according to the characteristics of wearable computer, Combined with the purpose of environment awareness, the frame structure of the whole system and the function of each module are determined, and the design of this system adopts the C graphical user interface application program QT, which is cross-platform, to realize the development of each module function. Finally, the actual use of the system is tested, the performance of the system is analyzed, and the desirable points of the system and the areas that need to be improved are summarized.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN912.34

【共引文献】