| 基于高阶谱的战场声目标特征提取 |
| |
| 引用本文： | 禹熊伟,张雄伟,陈功,沈刘平.基于高阶谱的战场声目标特征提取[J].弹箭与制导学报,2007,27(5):241-244,248. |
| |
| 作者姓名： | 禹熊伟 张雄伟 陈功 沈刘平 |
| |
| 作者单位： | 解放军理工大学通信工程学院,南京,210007 |
| |
| 摘 要： | 由于高阶谱对于高斯噪声不敏感的特性，在战场噪声环境中。提取被动声目标的双谱、双谱切片和相位耦合特征，通过建立声目标的高斯混合模型（GMM），进行声目标信号的识别。将基于此模型的识别结果与功率谱方法比较。仿真结果证明：基于高阶谱特征提取方法优于AR功率谱特征，该方法在抑制背景干扰中具有优越性。
|
| 关 键 词： | 特征提取 AR功率谱 双谱 相位耦合 |
| 收稿时间： | 2006-11-30 |
| 修稿时间： | 2007-01-23 |
The Feature Extraction from Battlefield Acoustic Targets Based on Higher-order Spectrum |
| |
| YU Xiong-wei,ZHANG Xiong-wei,CHEN Gong,SHEN Liu-ping.The Feature Extraction from Battlefield Acoustic Targets Based on Higher-order Spectrum[J].Journal of Projectiles Rockets Missiles and Guidance,2007,27(5):241-244,248. |
| |
| Authors: | YU Xiong-wei ZHANG Xiong-wei CHEN Gong SHEN Liu-ping |
| |
| Abstract: | The feature of the higher-order spectrum is insensitive to Gaussian noise. Based on it, the features of the passive acoustic signal, include bi-spectrum, the slice of bi-spectrum and phase-coupling etc, are extracted aiming to strong noise of battlefield multi-targets. An identification method based on GMM is proposed to identify every kinds of acoustic signal. Compared with AR spectrum, the influence of disturbed background-noise is reduced, and the clustering character of target feature is improved, which is proved by simulation results. |
| |
| Keywords: | feature extraction AR spectrum bi-spectrum phase-coupling |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
