基于加权SVM和m-χ分解的简缩极化SAR图像舰船检测

Ship detection in compact polarimetric SAR imagery based on weighted SVM and m-χ decomposition

摘要:

与全极化相比，简缩极化合成孔径雷达(SAR)因其更宽的幅宽，在海洋监视方面具有先天的优势。针对海上舰船目标检测，提出一种基于加权支持向量机(SVM)和m-χ分解的简缩极化SAR图像舰船检测方法。该方法首先对简缩极化的极化参数进行提取，构造加权特征向量，然后基于加权SVM分类器对简缩极化SAR图像舰船目标进行检测，最后利用m-χ分解后3个分量对应不同散射机制的差异进行虚警去除。基于NASA/JPL AIRSAR机载以及Radarsat-2星载全极化实测数据模拟的圆极化发射线极化接收(CTLR)模式的简缩极化数据实验结果表明，该方法能在舰船目标检测的同时，有效去除虚警和模糊噪声。

Abstract:

Compact polarimetric Synthetic Aperture Radar(SAR) has a congenital advantage in marine surveillance over full polarimetric SAR for its wider swath. A new ship detection method on compact polarimetric SAR image based on weighted Support Vector Machine(SVM) and m-χ decomposition is proposed. Firstly, the proposed method constructs the weighted feature vectors by extracting the compact polarimetric parameters. Then, the ship targets in compact polarimetric SAR image are detected by the weighted SVM classifier. Finally, the false alarms are wiped off according to scattering mechanism strength differences corresponding to the three components of m-χ decomposition. The NASA/JPL AIRSAR airborne full polarimetric data and the Radarsat-2 satellite-borne full polarimetric data are used to simulate the compact polarimetric data in the Circular Transmit-Linear Receive(CTLR) mode, and the experimental results show that the method performs well in detecting ship targets, and can remove the false alarms and ambiguities effectively.