Abstract: Aiming at the problem of the low accuracy of conventional airport main track detection method, this paper presents a method of the airport targets detection for Synthetic Aperture Radar(SAR) images. It uses Fuzzy C-Means(FCM) clustering to segment the airport. It utilizes morphology filtering and removing the connected domains whose number of pixels is less than the threshold to remove some interfering targets in the background. Then it uses Hough transform to detect lines in the edge image. The postprocessing is based on the characteristic that the runways of the airport are parallel pairs and the fusion method. Experimental result indicates that the algorithm can detect the runways of the airport successfully, and it has significant practical value.

Key words: Synthetic Aperture Radar(SAR) image, airport detection, Fuzzy C-Means(FCM) algorithm, morphology filtering, Hough transform

摘要： 针对现有机场主跑道检测方法提取目标不明确的问题，提出一种从合成孔径雷达(SAR)图像中自动检测机场主跑道的方法。使用模糊C均值算法分割原始SAR图像，采用形态学方法进行去噪处理，通过去除较小连通区域的方法以消除虚警目标，运用Hough变换检测主跑道，并基于直线平行线的特性对主跑道进行后处理，得到的检测结果具有较高的正确率和较低的虚警率。

关键词: 合成孔径雷达图像, 机场检测, 模糊C均值算法, 形态学滤波, Hough变换

CLC Number: 

• TP39