| 组合核函数支持向量机在水中目标识别中的应用 |
| |
| 引用本文： | 陆阳,王海燕,田娜.组合核函数支持向量机在水中目标识别中的应用[J].声学技术,2005,24(3):144-147. |
| |
| 作者姓名： | 陆阳 王海燕 田娜 |
| |
| 作者单位： | 西北工业大学航海学院,西安,710072 |
| |
| 摘 要： | 论文研究了支持向量机核函数构成条件以及不同核函数的特性，结合水中目标识别技术特点，提出了一种组合核函数支持向量机的方法。提取了基于小波变换的舰船辐射噪声奇异性、尺度-过零、尺度-能量特征，对水中目标进行了SVM分类识别。研究表明，基于组合核函数的支持向量机分类识别效果优于单独核函数的支持向量机识别效果。
|
| 关 键 词： | 支持向量机 核函数 目标识别 |
| 文章编号： | 1000-3630-（2005）03-0144-04 |
| 收稿时间： | 2004-05-25 |
| 修稿时间： | 2004-05-252004-07-11 |
a support vector machine with a hybrid kernel and its application in underwater target recognition |
| |
| LU Yang,WANG Hai-yan and TIAN Na.a support vector machine with a hybrid kernel and its application in underwater target recognition[J].Technical Acoustics,2005,24(3):144-147. |
| |
| Authors: | LU Yang WANG Hai-yan and TIAN Na |
| |
| Affiliation: | College of Marine Engineering, Northwestern Polytechnical University, Xi''an 710072, China;College of Marine Engineering, Northwestern Polytechnical University, Xi''an 710072, China;College of Marine Engineering, Northwestern Polytechnical University, Xi''an 710072, China |
| |
| Abstract: | The formation conditions and the characteristics of different kernel functions are studied. Based on the research of underwater target recognition, a method for training support vector machine with a hybrid kernel is proposed. After extracting characteristics of an underwater target such as Lipschitz singularity, scale zero-cross density and scale energy factor by using wavelet transformation, the hybrid kernel function SVM is used to recognize different targets. Experimental results indicate that using the hybrid kernel can give better performance compared to a single common kernel. |
| |
| Keywords: | support vector machine kernel function target recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《声学技术》浏览原始摘要信息 |
|
点击此处可从《声学技术》下载全文 |
|
