基于数学形态学和Hough变换的高分辨率遥感影像道路提取

城市交通是经济社会发展的重要推动因素和枢纽,因此道路信息的提取与更新显得尤为重要。利用遥感图像来提取道路信息已经成为近年来道路提取的主要方法,但高分辨率遥感影像的快速发展与应用,在提供丰富信息的同时使道路提取变得更加困难。针对目前基于高分辨率遥感影像提取道路存在的问题,提出一种基于数学形态学和Hough变换的道路自动化提取方法。基于Hough变换确定道路的走向,在此基础上选定线性结构元素,利用数学形态学进行道路提取。分别选择IKONOS影像和Quickbird影像验证本文提出的模型。实验结果表明,本文提出的模型能有效地提取弯曲型道路与直线型道路,取得了较好的结果。

Road Extraction Using High Resolution Remote Sensing Image Based on Mathematical Morphology

Urban traffic is the important driving factors for economic and social development. Therefore, it is particularly of importance to extract and update road information. In recent years, the road information extracted based on satellite images has become a main method of road extraction. However, with rapid development and application, high resolution satellite image not only provides rich context, but also makes the road extraction more difficult. Aiming at the problem of road extraction based on high resolution satellite image, a method for automatic extraction of road based on mathematical morphology and Hough transform is presented in this paper. First, the road direction based on Hough transform for selecting the linear structural element is determined and then the mathematical morphology for road extraction is adopt. The proposed model using IKONOS images and Quickbird images are validated. The experimental results show that the model proposed in this paper can effectively extract the curved road and the straight road, and obtain good results.