多特征、多方法融合的高分辨率影像道路网提取  

Extracting road networks from high-resolution remote sensing images using multi features and methods

作　　者：李润生 曹帆之 曹闻[1] 王淑香 LI Runsheng;CAO Fanzhi;CAO Wen;WANG Shuxiang(Data and Target Engineering College,Information Engineering University,Zhengzhou 450001,China)

机构地区：[1]信息工程大学数据与目标工程学院,郑州450001

出　　处：《国土资源遥感》2018年第3期33-39,共7页Remote Sensing for Land & Resources

基　　金：地理信息工程国家重点实验室开放基金"高分辨率遥感影像道路提取技术研究"(编号:SKLGIE2016-Z-3-2)资助

摘　　要：高分辨率影像上道路表现为宽度近似不变的条带状同质区域。根据此特征,提出了一种融合多特征、多方法的高分辨率影像道路网自动提取方法。该算法首先采用均值漂移聚类对图像稳态点图进行分类;然后运用Gabor滤波及张量编码,以线性显著性最大为准则识别道路中心点类;最后,运用张量投票和连通成分分析完成道路段连接及道路网组网。试验结果表明该方法能够准确、完整地提取高分辨率影像上道路网,提取的完备性和准确度优于对比算法。Roads on the high-resolution remote sensing images perform the stripe homogeneous region with ribbon-like shape and approximate width.According to these features,this paper presents a simple yet effective method of delineating road networks from high-resolution remote sensing images,which combines multi features and methods.The proposed method consists of three main steps.First,the mean shift algorithm is utilized to detect the modes of density of image points in spectral-spatial space which contain potential road center points and then detected mode points are classified into different classes by mean shift-based clustering on the basis of spectral information.Next,the combination of Gabor filtering and tensor encoding is used to identify the road class and to extract road center points.Lastly,road network is generated from detected road center points by means of tensor voting and connected component analysis.The experimental results demonstrate good performances of the proposed method in road network extraction,much better than the method proposed by Miao et al.

关 键 词：高分辨率 遥感影像 同质性 均值漂移 道路网组网 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]