高分辨率遥感影像主干道路提取的感知编组方法

Automatic extraction of main roads from high resolution remote sensing imagery based on perceptual organization

针对高分辨率遥感影像的道路提取受"同物异谱、同谱异物"干扰的问题,该文提出一种基于感知编组的高分辨率遥感影像主干道路自动提取方法:首先,使用Line Segment Detector算法提取影像中直线段信息。然后,利用高分辨率遥感影像上的道路几何特征对直线段进行感知编组;最后,经过长度约束得到道路信息。使用两景影像进行了主干道路提取实验。实验结果表明,两个实验中主干道路提取的完整率、正确率和检测质量都在96%以上。

Roads extraction from high resolution remote sensing imagery is still faced with lots of chal- lenges, especially the one ＂same object with different spectra, different objects with same spectrum＂ . We present a method for automatic extraction of main roads from high resolution remote sensing imagery in this paper. The proposed method is composed of three parts. Firstly, the algorithm, Line Segment Detec- tor, is used to extract the line segments from the images. Secondly, perceptual organization is realized to cluster the line segments based on the geometric characteristics of the roads. Finally, the road information is obtained through length constraint. Two images are used for testing the above proposed method. The ex- perimental results suggest that the completeness, correctness and quality of our proposed method on the two images are all larger than 96%.