基于多方向突堤扫描的极化SAR图像人工港口检测

发布时间：2018-03-25 20:37

本文选题：合成孔径雷达　切入点：港口检测　出处：《系统工程与电子技术》2017年02期

【摘要】：针对现有合成孔径雷达(synthetic aperture radar,SAR)图像中基于海陆两区域分割和单方向突堤扫描港口检测方法存在虚警率高的问题,提出一种基于多区域分割和多方向突堤扫描的人工港口检测方法。该方法首先使用基于极化SAR相干矩阵统计特性的多区域水平集分割方法将地物分割为水域、强散射建筑和其他陆地3类;然后在水陆分割结果上从多对正交方向对沿岸突堤进行扫描,从而确定出港口感兴趣区域;最后利用强散射建筑分割结果实现感兴趣区域中人工港口和虚警天然凸起地形的区分。基于RADARSAT-2对新加坡和湛江地区实际观测的极化SAR数据进行实验,结果表明本文方法较好地实现了沿岸人工港口的正确检测。
[Abstract]:In view of the problem of high false alarm rate in the existing SAR synthetic aperture radar radar SAR images based on the segmentation of sea and land regions and the detection method of single directional breakwater scanning port, An artificial port detection method based on multi-region segmentation and multi-directional embankment scanning is proposed. Firstly, a multi-region horizontal set segmentation method based on the statistical properties of polarimetric SAR coherence matrix is used to divide the ground objects into waters. Strong scattering buildings and other land types, and then scan the coastal breakwater from multiple orthogonal directions on the results of land and water segmentation, and then determine the area of interest to the port; Finally, the artificial port and false alarm natural raised terrain in the region of interest are distinguished by using the segmentation results of the strong scattering building. Based on RADARSAT-2, the polarimetric SAR data observed in Singapore and Zhanjiang are tested. The results show that the method can realize the correct detection of artificial ports along the coast.
【作者单位】：清华大学电子工程系;国家卫星海洋应用中心;
【基金】：国家自然科学基金(61490693) 民用高分重大专项(41-Y20A14-9001-15/16)资助课题
【分类号】：TN957.52

【参考文献】