摘要
为解决目前婴儿智能监护产品对婴儿哭声识别不准确的问题,设计了能准确识别婴儿哭声的婴儿智能监护系统;使用RT5350芯片作为处理器,使用PAP7501芯片实现对视频信号和音频信号的采集;对系统移植了Linux内核,使用多线程编程技术编写了应用程序;通过提取梅尔频率倒谱系数作为特征参数,使用动态时间规整算法作为识别算法,实现了对婴儿哭声的准确识别;该系统可以与云服务器通信,实现婴儿哭声报警功能;该系统可以与用户的终端设备通信,实现视频数据和音频数据的传输;该智能监护系统极大地方便了用户对婴儿的监护。
To solve the problem of the current infants＇ intelligent monitoring products＇ low recognition rates for the infants＇ crying,an infants＇ intelligent monitoring system which can recognize the infants＇ crying accurately was designed.The RT5350 chip was used as processor,the PAP7501 chip was used to capture the video signal and audio signal.A Linux kernel was ported to the system,the multi-threaded programming techniques was used to program the application programs.The Mel frequency cepstrum coefficient （MFCC） was extracted as the characteristic parameters,the dynamic time warping （DTW） algorithm was used as the recognition algorithm.The system can communicate with the cloud server to realize the infants＇ crying alarm function,and it can also communicate with the user＇s terminal device to realize the transmission of video data and audio data.The users can monitor the infants effectively with the help of the system.
出处
《计算机测量与控制》
2016年第7期105-108,共4页
Computer Measurement &Control
基金
天津科技支撑计划项目(14ZCZDGX00003)
天津市海洋经济创新发展区域示范项目(2015120024000473)
天津市自然科学基金一般项目(13JCYBJC18000)
作者简介
杜仲平（1990-）,男,吉林通化人,硕士研究生,主要从事嵌入式系统方向的研究.
