结合尺度空间的面向对象高分辨率影像城市道路提取

基于遥感影像的城市道路提取对于城市建设、规划和地图更新等有重要意义。针对高分辨率遥感影像城市道路网的复杂性,结合尺度空间思想提出一种面向对象的城市道路自动提取算法。在此基础上,使用Canny算子获取像元簇梯度图,并进行标记分水岭分割得到区域对象;建立城市道路与几何、光谱特征相关的道路规则,从分割结果中筛选出道路区域对象;使用形态学方法提取道路区域的骨架,并对骨架进行连接、光滑等后处理,最后输出道路网提取结果。实验结果表明,该方法用于复杂城市道路的高精度自动提取,对城市道路网更新有一定参考意义。

Object-based urban road extraction from high resolution imagery with space-scale theory

Urban road extraction from remote sensing image is significant for city construction ,planning and map updating .An automatic extraction algorithm of urban road by integrating scale‐space and object‐based approach is proposed in this paper .Furthermore ,canny operator is used to produce the gradient image ,where marker‐based watershed transform is conducted to achieve image regional objects .Rules are designed to choose the road objects step by step and other objects are discarded , based on the typical geometric and spectrum features of urban roads and the former cluster results .Then ,the morphological skeleton algorithm is used to generate centerline of these road objects followed by some post‐processing work ,such as skeleton smoothing and connection .The experiment results show that the proposed method is suitable for automatic extraction of urban road ,and the applicability of algorithm and the accuracy of results go for the updating of urban road‐network .