摘要
针对舱音译码系统“辨听”的不足,提出了基于计算机音频原理和专用软件的改进性舱音译码系统的舱音中话语声和开关声的声频特征获取方法;然后根据话语声和开关声是含有噪声的非平稳信号的特点,并依据小波变换理论提出了一维离散二进制Daubechies小波函数db1三尺度小波分解方法,对话语声和开关声降噪并进行深入处理,得到频率—能量幅值等更多信息,结果有利于掌握飞机飞行状态,调查飞机事故原因,也为分析同类舱音信号提供了参考依据。
In order to solve the cockpit voice decoding system＇s ＂differentiating ＆ hearing＂ shortage,based on audio frequency principle of ameliorated cockpit voice decoding system and special software,the methods of obtaining the spectrum characteristics of the voice and switch sound,recorded by airplane Cockpit Voice Recorder（CVR）,is put forward. Because the voice and switch sound are no steady-state signals and there are noise in them,therefore,another method, one-dimensional discrete Daubechies wavelet function dbl three scales decomposition method through wavelet transformation ,is introduced ,and the voice and switch sound are decomposed and analyzed again.The noise is removed, more information including frequency-energy magnitude is obtained.All the results are favorable to gain airplane flight state data and investigate airplane wreckage causes.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第25期187-189,196,共4页
Computer Engineering and Applications
基金
国家自然科学基金资助项目(编号:60572182)
中国民航总局科技项目
关键词
计算机
飞机
舱音记录器
小波分析
computer, airplane, cockpit voice recorder, wavelet analyses
