融合高分影像和LiDAR数据的城市道路提取

为避免由于城市道路复杂及树木建筑的阴影遮挡导致从遥感影像中提取道路信息不准确的问题，本文采用高分影像和LiDAR数据相融合的方法实现城市道路的提取，并使用一种基于最小面积外接矩形（MABR）的后处理改进方法进行完善。首先对试验区进行数据配准；然后应用FNEA算法进行图像分割，并使用随机森林分类法进行分类，将影像融合和对象形状指数等相关算子应用到道路提取中；最后去除植被和建筑物，完善道路填充，提取出道路完整信息。结果多伦多和台安试验区的道路完整度分别为95.41%和90.84%，准确度分别为83.07%和85.63%。本文方法可有效去除伪道路信息，提高道路提取完整度，较好地实现了道路信息提取。

Urban road extraction with fusion of high-resolution images and LiDAR data

In order to avoid the problem of inaccurate road information extraction from remote sensing images due to the complexity of urban roads and the shadow occlusion of trees and buildings. The method of fusion of high-resolution images and LiDAR data is used to achieve the extraction of urban roads, and a post-processing improvement method based on the minimum area boundary rectangle (MABR) is proposed to improve the extraction. First, we realize the data registration in the area. Then, we apply the FNEA algorithm for image segmentation and use the random forest classification method for classification, apply image fusion and object shape index and other related operators to the road extraction. Finally we extract complete road information by removing the vegetation and buildings and improving the road filling. The results show that the road integrity of Toronto and Tai’an Study area is 95.41% and 90.84%, and the accuracy is 83.07% and 85.63%. This method can effectively remove false road information, improve the completeness of road extraction, and achieve better road information extraction.