基于HPF融合与数学形态学的ETM＋遥感影像道路提取

摘要

道路作为构成城市格局的骨架,是城市发展过程中更新最快的要素之一,因此,基于数学形态学的ETM＋遥感影像的道路信息提取具有重要的现实意义.以合肥市某一地区的ETM＋影像为例,提出一种基于HPF融合与数学形态学的方法.首先对原始影像进行预处理,对原始影像的多波段与全色波段影像进行HPF融合,预处理结果得到的RGB图像转换为灰度图像,再转换为二值化图像,然后根据道路的线性特征,运用数学形态学对图像进行去噪（独立斑块及粘连斑块两部分）、膨胀腐蚀、开运算、细化等运算处理,应用此方法能够较完整的得到道路提取的结果,并对该方法进行了验证.实验结果证明：该方法可以较准确、完整地从分辨率遥感影像中提取主干道路网络,剔除非道路地物的影响,具有简单、实用性特点.%As a framework of the city pattern,the road is one of the fastest elements in the process of city development.Therefore,to extract road information based on mathematical morphology from ETM＋remote sensing image is of important and realistic significance.With an area of Hefei ETM＋image as an example,this paper proposes a method based on HPF fusion and mathematical morphology.Firstly,pretreat the original image,HPF fuse multi-spectrum bands and full-color band of the original image,convert pretreatment result of RGB image into gray image,the gray image to the binary image,and then based on the linear characteristics of the road,make use of the mathematical morphology to denoise the image（independent patch and adhesion plaques）,inflat and corrode,open operation,refine the image,thus get the result of road extraction.Experimental results show that this method can be used to accurately extract trunk road network from remote sensing image resolution and eliminate the influence of the road features,which are simple and practical.