合成孔径雷达图像中机场跑道的自动识别

为解决雷达图像中的机场跑道识别问题 ,提出了一种合成孔径雷达图像中机场跑道的自动识别方法 .首先 ,对原始图像滤波并提取图像轮廓 ,其次对轮廓图作Radon变换 ,保留最大的几个Radon变换系数 ,将其余系数置 0 ,并对Radon逆变换得到的图像进行二值处理 ,从而过滤掉大部分干扰线条 ,仅保留图像中的主要直线条 ,最后利用直线模板滤除残余干扰线条 ,并根据阈值检测出机场跑道 .实验结果表明 ,所提方法能够准确地识别出复杂背景下的水平、垂直和倾斜的机场跑道

Automatic Recognition of Airfield Runway in Synthetic Aperture Radar Images

An approach of automatic airfield runway recognition in a SAR(synthetic aperture radar)image was proposed. In the proposed method, an the original image was filtered and transformed into binary one for contour extracting, Radon transformation was performed on the contour image and several maximum Radon transformation coefficients were retained while the others were set to zero, and then Radon inverse transformation was applied and the image was transformed into binary one, thus the most of the noisy lines could be filtered and only major straight lines in the image were retained; a line template was used to remove remaining noisy lines; airfield runways were picked out according to a threshold. Experimental results show that the proposed approach can accurately recognize horizontal, vertical and inclined runways of an airfield in a complicated background.