摘要
根据直升机信号在更多情况下表现为非高斯的随机过程，提出采用高阶统计量对直升机声信号进行建模与特征提取，导出基于三阶累积量的直升机声信号非高斯参数双谱模型，讨论了双频域内直升机声信号特征提取的方法，并结合实例给出说明。最后指出将高阶统计量分析方法用于直升机声目标识别具有十分有意义的应用前景。
Based on the fact that the acoustic signals of a helicopten behave as non Gaussian characteristrics in most cases, a signal processing method using higher order statistics is proposed modelling and extracting the featrues of the acoustic signals of the helicopter. A Non Gaussian parametric bispectral model is derived via third order cumulants and the method for extracting the features of helicopter signals in bi frequency domain is also discussed by using an example. Finally, this paper points out that it is significant to improve anti noise performance of its acoustic recognition system by using higher order statistics to process helicopter acoustic signals.
出处
《探测与控制学报》
CSCD
1999年第2期27-31,共5页
Journal of Detection & Control
关键词
双谱估计
直升机
声信号特征
声目标
目标识别
higher order statistics
parametric model
bispectral estimation
helicopter acoustic signal
feature extraction
