基于高分辨率遥感图像的道路提取研究

道路是现代交通的主要组成部分,对于管理和更新地理信息系统数据库中的道路信息非常重要.目前,自动提取道路网络的主要数据源为遥感图像数据,但随着近年来遥感影像的地面分辨率不断提高,图像中地物信息愈加丰富,对图像中道路信息的提取难度也随之增大.文章主要展开一种利用机器学习对高分辨率遥感图像的道路提取研究.首先对高分辨遥感图进...

Research of Road Extraction from High Resolution Remote Sensing Images

The road is the main part of modern transportation,which is very important for managing and updating the road information in the global information system database. At present,the main data source of automatic extraction of road networks is in remote sensing image data. However,with the continuous improvement of ground resolution of remote sensing images in recent years,the ground object information in the image is more and more abundant,and the difficulty of extracting road information in the image is increasing. This paper mainly focuses on the research of road extraction from high-resolution remote sensing images using machine learning technique. Firstly,the high-resolution remote sensing image is preprocessed,then the image features are extracted,and the features are trained by BP neural networks. Finally,the high-resolution remote sensing image that needs road extraction is segmented. When each region is detected,the non-road region in the image is removed,and the road information in the region is extracted by this morphological method. Experiments show that the efficiency of this method is significantly improved when dealing with the interference of buildings and vegetation on road extraction.