一种基于区域生长与空间形状约束的高分辨率遥感图像道路提取方法  

A Road Extraction Method from High-resolution Remote Sensing Image Based on the Regional Growth and Space Shape Constraints

作　　者：叶勤[1,2] 张小虎[1,2] 王栋[1] 

机构地区：[1]同济大学测量与国土信息工程系,上海200092 [2]同济大学遥感与空间信息技术研究中心,上海200092

出　　处：《遥感信息》2010年第2期25-29,共5页Remote Sensing Information

基　　金：上海市科委项目"生态城镇建设管理实证研究"课题支持(课题编号:08DZ1205708)

摘　　要：从遥感图像上快速、高效地提取道路等地理信息一直是遥感数据处理的研究热点,由于高分辨率影像的特点,仅使用光谱信息或形态特征进行道路的提取都难以得到满意的结果。本文将区域生长与空间形状约束相结合进行道路提取:首先利用道路种子点处的光谱信息进行道路区域的生长,提取光谱信息一致的道路区域,在提取出的道路条状区域的基础上,根据道路具有的形状特点,利用形态学滤波、面积滤波等方法进行高分辨率图像上的道路边线信息的确定,得到精确的道路数据。文章对高分辨率航空遥感影像进行了实验验证了该方法的有效性。High effective extraction of geographic information,such as roads,is a research hotspot in remote sensing data processing.Since the characteristics of high-resolution images,the extraction of road can not obtain satisfactory results with only spectral information or morphological characteristics.In this paper,regional growth and space shape constraints were combined to carry out the road extraction： first,using the spectral information of seed points to conduct the road＇s region growth and extract road regions with same spectral information.Then,on the basis of the extracted road strips,in accordance with the shape of the characteristics of the road,with the use of morphological filtering and the area method of filtering,the edge of the road in high-resolution image were identified.Finally,precise road data were derived.Experiments with high-resolution remote sensing image were conducted to verify the effectiveness of the method in the paper.

关 键 词：区域生长 形态学滤波 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]