基于结构特征分析的COSMO-SkyMed图像商用船舶分类算法

发布时间：2018-03-11 06:29

本文选题：合成孔径雷达　切入点：特征提取　出处：《遥感技术与应用》2014年04期 　论文类型：期刊论文

【摘要】：船舶分类与识别对于海洋交通运输监测与管理具有重要意义,同时也是SAR海洋应用的重要组成部分。COSMO-SkyMed高分辨率合成孔径雷达(SAR)图像下,商用船舶的结构轮廓明显,散货船、集装箱船和油船的特征清晰可辨,为船舶识别分类提供有效支持。提出了一种基于结构特征分析的商用船舶分类算法,通过提取核密度估计值、船舶积分主轴位置及左中右3部分积分量比例等特征,可实现船舶类型的区分。通过在东海试验区的同步实验,证明COSMO-SkyMed图像商用船舶分类算法的平均分类精度达到89.94%。
[Abstract]:Ship classification and recognition is of great significance for ocean traffic monitoring and management, and is also an important part of SAR ocean application. In the image of COSMO-SkyMed high resolution synthetic aperture radar (sare), the structure of commercial ships is obvious, and the bulk carrier is very important. The features of container ship and tanker are clearly distinguishable, which provides effective support for ship identification and classification. A classification algorithm for commercial ships based on structural feature analysis is proposed. The kernel density estimation is extracted. The ship type can be distinguished by the position of the ship integral spindle and the proportion of the integral quantity between the left, middle and right parts. Through the synchronous experiment in the East China Sea experimental area, it is proved that the average classification accuracy of the commercial ship classification algorithm based on COSMO-SkyMed images reaches 89.94.
【作者单位】：中国科学院遥感与数字地球研究所数字地球实验室;中国科学院大学;
【基金】：国家自然科学基金项目(41331176)
【分类号】：U675.7;TN957.52

【相似文献】