摘要
针对高分辨率遥感影像的特点,提出了一种基于面向对象思想的自动道路提取方法.首先对遥感影像进行双边滤波,平滑细节信息并保留道路边缘信息;然后使用模糊C均值算法分割图像以得到独立的地物对象,并结合道路几何特征对各对象滤波得到候选道路段;使用区域增长算法形成道路网,最后使用形态学方法实现道路网的修整和细化.实验表明,该方法无需人工选取道路种子点,就可以在不同场景的遥感影像中有效地提取出道路目标.
An automatic object oriented method of road extraction is proposed for high-resolution remote sensing image.At first,the bilateral filter is used in image to smooth the detail information and retain road edge.Then the image is segmented to independent ground objects by FCM algorithm and each object is filtered combining geometric feature to obtain candidate road segments.The region growing algorithm is used for road segment connection to get network.Finally,the road network is repaired and thinned by the morphology method in post processes.The experiments show that the method can extract the road target efficiently from the high resolution image in different scenes without selecting road seeds artificially.
出处
《兰州交通大学学报》
CAS
2017年第1期57-61,共5页
Journal of Lanzhou Jiaotong University
基金
国家自然科学基金(60962004
61162016)
甘肃省青年科技基金计划(148RJYA011)
兰州交通大学青年基金(2015003)