结合形态学理论与Hough变换的SAR图像线目标检测方法

针对典型线状军事目标,提出一种SAR图像的目标检测方法。该方法首先将分块阈值分割的思想应用于未经斑点噪声抑制的原始SAR图像,得到ROI(Region of Interest)图像;然后,利用区域的几何特征有效地剔除大量虚警,并采用形态学梯度算子提取目标的边缘信息,与传统的Canny边缘检测相比,边缘轮廓更加连贯;最后,利用Hough变换对梯度图像进行直线检测,得到机场跑道的边缘。该方法对原始SAR图像采用传统的图像处理技术进行目标检测,因此,比基于SAR图像统计特性的目标检测方法简单易行。对真实SAR图像的实验结果验证了该方法的有效性。

A Linetype Target Detecting Method for SAR Images Based on the Combination of Morphology and Hough Transform

This paper proposes a target detecting method for typical linetype man-made targets in SAR images.First,the local threshold method is applied to the SAR image in which the intrinsic speckle noise has not been depressed,and the region of interest(ROI) is obtained.Secondly,the geometrical characteristics are utilized in order to reduce false alarms.Thirdly,the edge information of the target is extracted with the basic morphological gradient detector instead of the traditional Canny detector,which can get the edges with better consistency.Finally,the Hough transform is used to detect straight lines in edge images automatically.Owing to the traditional image processing techniques,this method can be realized more easily and provided with more robustness,comparing with other detection methods,such as those based on statistical models of SAR images.The experiments are carried out on real SAR images and the results demonstrate the validity of the proposed method.