论文部分内容阅读
研究用短波语音通话携带的飞机舱室噪声对飞机类型进行识别的方法。分析了飞机舱室内噪声在短波信道和语音通话干扰下的物理特性,定义了估计语音段的飞机噪声信噪比的公式,提出了自适应的抑制语音增强飞机噪声的模型,通过CZT变换分别提取目标信号不同频段的功率谱密度级特征,并设计了用支持向量机进行分类识别的二叉分类树。对8类现场实测数据进行实验:增强后语音段的平均信噪比提高约22 dB,分类树对语音应答间隔噪声、语音段信号和增强后的信号的平均识别率分别为82.79%,15.25%,50.18%。实验表明:应答间隔噪声可用于飞机类型识别;语音抑制算法带来较大的信噪比和识别率增益,证明语音段蕴含有助于飞机类型识别的重要信息,可为后续的研究奠定基础。
A method of identifying aircraft types using aircraft cabin noise carried by short-wave voice calls. The physical characteristics of aircraft cabin noise under the interference of short-wave channel and voice call are analyzed. The formula to estimate the signal-noise ratio of aircraft noise in speech segment is defined. An adaptive noise suppression model of speech enhancement aircraft is proposed and extracted by CZT transform The power spectral density of different frequency bands of the target signal, and design a binary classification tree using SVM for classification and recognition. The experimental results show that the average signal-to-noise ratio (SNR) of the enhanced speech segment increases by about 22 dB, and the average recognition rate of the classification tree to the speech response interval noise, speech segment signal and the enhanced signal are 82.79% and 15.25% , 50.18%. Experiments show that the response interval noise can be used for aircraft type identification. The speech suppression algorithm brings a large signal to noise ratio and recognition rate gain. It proves that the speech segment contains important information that helps identify the aircraft type, which can lay a foundation for subsequent research.