| 潘谢帆,周胜增,蒋小勇,杜选民.一种基于支持向量机分类的瞬态信号检测方法[J].声学技术,2011,(5):453~455 |
| 一种基于支持向量机分类的瞬态信号检测方法 |
| A method of transient signal detection based on the support vector machine |
| 投稿时间：2010-10-15 修订日期：2010-01-13 |
| DOI： |
| 中文关键词: 瞬态检测 支持向量机 时频细胞 高斯核 |
| 英文关键词: transient detection support vector machine time-frequency cell Gaussian kernel |
| 基金项目: |
|
| 摘要点击次数: 1296 |
| 全文下载次数: 817 |
| 中文摘要: |
| 用支持向量机(Support Vector Machine,SVM)方法对水下运动目标辐射噪声的谱图进行高维空间下的最优划分,实现水下瞬态信号的有效检测。其基本思想是将时频谱图拆分成若干时频细胞单元(Time Frequency Cell,TFC),选择合适的高斯核向量机,寻找时频细胞单元间的差异性,进而实现对瞬态信号的检测。海试数据处理表明该方法检测瞬态信号的有效性,运算量小且稳健性高;与常规能量检测方法相比,更易确定检测门限,减少虚警。 |
| 英文摘要: |
| A novel method of transient signal detection based on the support vector machine is proposed in the paper,in which the decision boundary is obtained from dissimilarity between the transients and noise,calculated from the Gaussian kernel solution in high-sphere,which is based on an analysis of the time-frequency cells of segmented spectrograms.The sea tests show that the method can detect the transients in real time with its robustness and small computation.Compared with the power detection,the proposed method can reduce the false alarm due to confirming the detection threshold easier. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
