摘要
在高分辨率航空或卫星图像中提取道路特征具有越来越重要的实际意义。道路与背景通常存在灰度差异,论文首先求取高斯平滑图像的偏导数,进而确定图像中的脊点,然后根据连接准则连线,再构造判别规则删除道路虚警,最后利用Canny边缘图修正道路中心线。算法还对交点区域进行了有效处理。实验证明,该方法在一定程度上能够消除树、建筑物阴影和汽车对道路提取造成的影响,提取的道路中心线比较准确,能为道路网检测的后续处理提供重要的数据支持。
Road extraction from high-resolution aerial or satellite images shows increasing significance.Usually,there is difference in the brightness of road and background.Hereby,the salient points can be gotten by the partial derivatives of the Gaussian smoothed image and are connected to lines,then the false roads are eliminated by constructing criterion function.Finally,the coordinates of the road centerlines are modified by using Canny operator.The algorithm also includes the processing of the road in the junctions.The experiment results show this method can remove the effects of the shadows of trees and buildings as well as motorcars and the accuracy is higher.It can provide the important data support for the further treatment of the road network detection.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第30期185-187,217,共4页
Computer Engineering and Applications