一种遥感图像机场跑道的多尺度提取方法

遥感图像机场跑道边缘的提取是机场识别的主要方法。传统Hough变换在线段提取方面具有较高抗噪性,但用于遥感图像机场跑道提取时存在边缘定位性较差及弯曲跑道误检率高的问题。本文提出了图像空间多尺度Hough变换方法,提高了Hough变换在提取、检测线段时的定位能力;将传统Hough变换的对参数空间改进为相对参数空间,增强了Hough变换检测小线段的性能,并利用分段线段的连接,达到机场跑道边缘中直线段与曲线线段检测的目的。试验结果表明该方法在保持Hough变换高抗噪性的同时,可有效地检测复杂背景下遥感图像中的机场跑道边缘,并保证了边缘的连接性。

A method of extracting airfield runway in remote sensing image

Edge extractin of airfield runway in remote sensing image is very important to airfield recognition. Multiscale Hough transform combining Hough transforrn and the partition of image space, solves the problem of location of line segments. Relative parameter Hough transform improves the ability to detect short line segments. The straight and curve lines of airfield runway edges can be extracted by connecting line segments of different lengths. The experimental results of remote sensing images show that the proposed method can accurately detect the edges and has robustness in the presence of noise.