复杂SAR场景中机场跑道的提取

Extraction of Runways in Complex Synthetic Aperture Radar Images

本文提出了一种从大幅SAR场景图像中提取机场跑道的方法。跑道是机场的最明显的特征,在图像上可抽象为相隔一定距离,具有一定长度的平行线对。算法首先检测图像边缘,然后连接成线段,从中搜索平行线对,最后进行验证。算法的特点在于连接线段时,通过将图像域中的线段转化成极坐标中的点,将线段连接问题转化为点的聚类问题,并利用贝叶斯估计原理构造相似性测度准则函数,利用区域生长聚类方法将断裂的但属于同一条直线的点聚集起来。实验结果表明,该算法能够从大幅SAR图像中提取机场跑道,具有重要的实际应用价值。

A method for extracting runways in large SAR images is put forward by this paper since the runways are the most obvious characteristic of the airport,which has two parallel straight lines along their edges. A series of process is included in the method：detecting edges, linking them into line segments, then searching anti -parallels, and finally verifying the extraction result. The specialty of the algorithm lies in the phase of searching the anti-parallels. The line segments in Cartesian coordinates are transformed into the points in polar coordinates. Based on the characteristic that fragments on a same line have same polar coordinates, the problem of line linking is turned into that of point clustering. And at the same time,we construct a metric function by means of Bayesian Estimation for measuring whether two points are close enough or not to be clustered. Experimental result indicates that the algorithm can extract runways in large complex SAR images, and so, has significant practical value.