基于极化SAR图像分类的海上舰船检测

本文针对极化熵类检测方法的不足,在极化特征分解以及Touzi等人工作的基础上,提出了能够更全面的表征舰船和杂波差别的特征矢量,并提出了一种基于特征矢量的非监督分类方法。使用该方法进行海上舰船的检测,不仅取得了很好的舰船和海面的分离效果,而且也得到了较好的舰船与其他人造目标的区分效果。实测数据的检测结果证明该分类方法具有很好的收敛性,是一种有效的舰船检测方法。

Ship detection based on classification of polarimetric SAR images

In this paper,we propose the feature vector based on eigen decomposition and Touzi's work,which is more effective to signify the difference between ships and ocean or other man-made objects than the method based on entropy.Moreover,we give a new method of unsupervised classification based on feature vector.The result of using this method for ship detection indicates that it can not only distinguish the ships from ocean but also can discriminate ships from other man-made objects and the classifier is an effective and convergent method of ship detection.