融合机载雷达和航空正射影像的道路提取  

Fusion of airborne LiDAR and aerial DOM for road extraction

作　　者：曾妮红 岳迎春[1,2] 黄迟[1] 

机构地区：[1]中国地质大学(武汉)信息工程学院,武汉430074 [2]武汉工程科技学院,武汉430299

出　　处：《测绘科学》2016年第5期96-99,151,共5页Science of Surveying and Mapping

基　　金："中央高校优秀青年教师专项资金"资助项目(2013-2015)(CUGL130257)

摘　　要：针对单一遥感数据源提取道路信息中存在的问题,该文提出一种融合多源数据的方法:采用面向对象思想,融合机载激光雷达和航空正射影像,运用对象平均强度指数进行道路精确提取。首先利用点云强度信息进行道路粗提取,再结合影像光谱信息对粗提取点云进行优化,得到"一次点云";然后通过分形网络演化算法对影像进行多尺度分割形成对象,将"一次点云"导入分割对象,利用对象平均强度辅之以对象面积完成道路精确提取。实验结果表明,该方法提取流程简单快速、无须精确配准、道路提取精度较高。Aiming at the problem in extracting roads with a single source of RS data,the paper presented an extraction method fusing multi-source data：the concept of the OAI index was put forward to extract road features by the fusion of LiDAR and aerial DOM based on object-oriented technique.The point cloud intensity information was first applied to extract rough roads,and the spectrum information of image was combined to optimize the coarse extracted point cloud;then fractal net evolution approach algorithm was used to divide the image into homogeneous region objects,and Once Point Cloud was imported to the multi-scale segmented objects;finally OAI and object area were used to extract road features accurately.Experimental result showed that the proposed method would be simply and rapidly extract roads with high precision,without accurate registration.

关 键 词：机载激光雷达 航空正射影像 面向对象 对象平均强度 道路提取 

分 类 号：P23[天文地球—摄影测量与遥感] TP75[天文地球—测绘科学与技术]