| 高阶统计量在水下目标识别中的应用 |
| |
| 引用本文： | 程广涛,戴卫国,李茂宽. 高阶统计量在水下目标识别中的应用[J]. 青岛大学学报(自然科学版), 2004, 17(4): 66-69 |
| |
| 作者姓名： | 程广涛 戴卫国 李茂宽 |
| |
| 作者单位： | 海军潜艇学院水声中心,山东,青岛,266071 |
| |
| 摘 要： | 利用了高阶统计量进行特征提取的方法，着重研究了采用双谱分析的方法提取舰船辐射噪声，给出了双谱矩阵处理的峰值化算法，提取了47维的特征向量并利用BP神经网络对他们进行训练与识别，结果表明，A类目标的正确识别率为83％，B类目标为88％，C类目标为85．19％，D类目标为85．7％。
|
| 关 键 词： | 高阶统计量 舰船辐射噪声 识别 特征提取 |
| 文章编号： | 1006-1037(2004)04-0066-04 |
| 修稿时间： | 2004-06-17 |
Higher-order Statistics Analysis Applied to Target Recognition |
| |
| CHENG Guang-tao,DAI Wei-guo,LI Mao-kuan. Higher-order Statistics Analysis Applied to Target Recognition[J]. Journal of Qingdao University(Natural Science Edition), 2004, 17(4): 66-69 |
| |
| Authors: | CHENG Guang-tao DAI Wei-guo LI Mao-kuan |
| |
| Abstract: | In order to recognize correctly the signal of ship-radiated noise, the method of feature extraction based on Higher-order statistics is presented. Bispectrum analysis is adopted to the target recognition, and apex arithmetic is given. 47 dimension bispectrum features are extracted from different targets by bispec trum estimation . Based on the features, four types of ships are classified by means of back-propagation ar tificial neural network. The results show that the target recognition rates about four types of ships A,B,C and D are 83.3%,87.8% ,85.2% and 85.71%,respectively. |
| |
| Keywords: | Higher-order statistics ship-radiated noise target recognition feature extraction |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
