基于PCA算法的SAR图像舰船目标长宽特征提取

Ship's Feature Extraction from SAR Image based on PCA

随着SAR图像成像技术的不断发展,几何特征被广泛应用在目标识别中,长宽特征因其简单直观、效率高、易于提取等优势,常被作为船只类型的初始判定,针对SAR图像舰船目标长宽特征提取问题,提出一种新的方法。首先通过水平集以及形态学方法获得预处理后的目标图像,利用PCA算法获取SAR图像舰船目标的长轴,结合最小二乘椭圆拟合方法获取舰船目标的短轴,最终得到舰船目标的长宽特征。通过实测SAR图像处理结果表明,该方法能够在背景杂波干扰下,抑制相干斑噪声的影响,提高了长宽提取的精度,是一种有效的舰船目标长宽特征提取方法。

In this paper, a length and width feature extraction method for a ship target in SAR images is proposed. Firstly, the SAR images are segmented by the level sets. The domain elimination method is adopted for the segmented images to remove clutter in the next step. Consequently, the slice image of target is obtained. Then, the PCA method is used to obtain the long axis of the ship in SAR image, and the ellipse fitting method is applied to get short axis of the ship target. Thus, the feature information of the ships including length and width is obtained. Experimental results illustrate that the proposed method can extract the length and width feature of a ship target in SAR images effectively and accurately. Meanwhile, it can weaken the influence of speckle noise and background clutter in SAR images.