基于极化雷达影像的海岛礁提取

提出了一种利用RADARSAT-2全极化SAR影像和极化特征参数提取精确的海岛礁范围的技术方法。极化特征熵参数描述了目标散射的随机性,与海水相比海岛礁处于较高的去极化状态,因此海岛礁的熵值明显大于海水的熵值。首先本文利用EM(Expectation Maximization,最大数学期望)算法自动计算的提取海岛礁最佳阈值对熵参数文件进行阈值分割,得到海岛礁的初始分割结果。由于受到船只和海水表面波浪的影响,海水部分也会存在与海岛礁近似的熵值。因此初步阈值分割得到的海岛礁结果会有部分海水和船只等,利用PSNR(Peak value signal-to-noise ratio,峰值信噪比)提取海水大致范围并剔除海水范围内初始分割结果中的噪声部分。最后根据TM影像提取的海岛礁范围进行精度评价,实验结果表明该技术方法能够从极化SAR影像上准确提取海岛礁范围。

Delineation of Sea Island Using Polarimetric SAR Image

This paper proposes a method of delineating the sea island in RADARSAT-2 polarimetric SAR image.The polarimetric parameter of entropy generated from Cloud＇s Decomposition Algorithm describes the randomness of targets＇ scattering.Compared with sea water,the islands show higher polarimetric entropy due to their strong depolarization effect.Therefore,the islands are initially extracted from the entropy value using a threshold,and the threshold value is determined with the Expectation Maximization（EM） method.But,vessels and sea waves may bring in a false alarm because they could also present a similarly high entropy value as island.To exclude these false alarms,this paper used PSNR（Peak Value Signal-to-Noise Ratio） to extract approximate range of the sea water,then using the range of sea water to eliminate the noise parts in the initial segmentation result.Finally,the sea island extracted result from TM image is used to evaluate the proposed method,and the experiment result shows that it can achieve the accurate extraction of the sea islands from the polarimetric SAR image.