| 赵亚楠,李钢虎,曾渊.基于最小均方无失真响应和支持向量机的被动声纳目标识别[J].声学技术,2011,(3):223~226 |
| 基于最小均方无失真响应和支持向量机的被动声纳目标识别 |
| A method of passive sonar target recognition based on Minimum Variance Distortionless Response and Support Vector Machine |
| 投稿时间：2010-11-02 修订日期：2011-02-03 |
| DOI： |
| 中文关键词: 最小均方无失真响应(MVDR) 被动声纳目标识别 支持向量机 |
| 英文关键词: MVDR passive sonar target recognition SVM |
| 基金项目: |
|
| 摘要点击次数: 1280 |
| 全文下载次数: 9 |
| 中文摘要: |
| 为了有效地进行被动声纳识别,研究了一种运用最小均方无失真响应(Minimum Variance Distortionless Response,MVDR)谱系数作为特征参数,用多分类支持向量机作为分类器,进行被动声纳目标识别的方法.实验表明,在不同数目的训练样本情况下,基于最小均方无失真响应谱系数和多分类支持向量机的被动声纳目标识别方法使系统的性能显著提高,具有很好的识别效果和应用价值.其优于传统的神经网络作为分类器的识别方法,尤其是在训练样本较少情况下,识别率具有很大的提高. |
| 英文摘要: |
| The method of passive sonar target recognition by using Minimum Variance Distortionless Response as characteristic parameter and Multi-classification Support Vector Machine as classifier is studied.Experiments results show that the method based on MVDR and Multi-classification SVM can improve the sonar performance greatly with good recognition effect and application value in the case of different training samples.It works better than the traditional method which takes neural network as classifier,and increases the rate of recognition greatly under the condition of less training samples. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
