| 基于BP神经网络的SAR影像水面舰船检测 |
| |
| 引用本文： | 宋玮,张杰,姬光荣.基于BP神经网络的SAR影像水面舰船检测[J].地球信息科学,2004,6(3):111-114. |
| |
| 作者姓名： | 宋玮 张杰 姬光荣 |
| |
| 作者单位： | 中国海洋大学,青岛,266071;国家海洋局第一海洋研究所,青岛,266061;国家海洋局第一海洋研究所,青岛,266061;中国海洋大学,青岛,266071 |
| |
| 基金项目： | 国家863课题海洋卫星遥感综合应用技术(2001AA634020) |
| |
| 摘 要： | 分析了水面舰船的SAR成像能力和探测优势。在总结已有的舰船检测方法的基础之上,根据舰船目标在SAR图像中的成像特点,提出了利用BP神经网络实现SAR水面舰船目标识别实验。进一步利用形态学方法对检测结果进行修整,结合已知的SAR图像的像元大小,提取舰船的长度和宽度信息,同时利用原始图像的经纬度内插实现舰船目标中心的定位。结果表明,BP神经网络具有很强的自适应性,在低信噪比的情况下,也能够取得良好的效果。
|
| 关 键 词： | 合成孔径雷达(SAR) 舰船检测 神经网络 反向传播(BP) |
| 收稿时间： | 2003-02-25; |
| 修稿时间： | 2003年2月25日 |
Ship Detection on SAR Images Based on BP Neural Network |
| |
| SONG Wei.Ship Detection on SAR Images Based on BP Neural Network[J].Geo-information Science,2004,6(3):111-114. |
| |
| Authors: | SONG Wei |
| |
| Institution: | 1. Ocean University of China, Qingdao 266071, China;
2. First Institute of Oceanography, State Oceanic Administration, Qingdao 266061, China |
| |
| Abstract: | The imaging abilities and detective advantages of synthetic aperture radar(SAR)are analyzed.According to SAR imaging feature to ships,a method using back propagation(BP)neural network to detect shipson SAR image is put forward on the basis of summarizing the existing ship detection methods.Then morphologicmethod is used to adjust the net detective result.And ship length and width are also extracted depending on theknown pixel size of the SAR image.At the same time,the ship center can be oriented by interpolation usinglatitude and longitude information of the source image.Finally,tests are carried out by this method,whoseresults indicate that the BP neural network has very strong self-adaptability.Although under the circumstancesof low Signal-to-Noise,this method can also get good result. |
| |
| Keywords: | synthetic aperture radar(SAR) ship detection neural network back propagation |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《地球信息科学》浏览原始摘要信息 |
| 点击此处可从《地球信息科学》下载免费的PDF全文 |
|
