基于无人机遥感影像拓扑分析的三维重建

Topology analysis-based 3D reconstruction from UAV images

在无需任何地面控制点或其它先验知识前提下,探索一种基于无人机遥感影像的三维重建方法。利用无人机飞控数据建立的影像拓扑结构,依次通过特征点提取、影像匹配、从运动恢复结构等步骤估计出相机位置和姿态参数,并恢复出场景特征点云信息,最后对重建精度进行分析。试验结果表明,文中提出的方法可快速、可靠地实现较高精度的三维模型重建。

It presents a low-cos t,automated 3D-reconstruction approach without a previous camera calibration or any other prior knowledge of the scene to generate 3D scene based on unmanned aerial vehicle system （UAVs） images.For scene reconstruction,the camera location and orientation parameters,as well as a sparse 3D point cloud through feature extraction,image matching and an iterative,robust Structure from Motion （SfM） procedure are recovered from digital sequence images based on image topology analysis with flight-control data acquired by the UAVs.Results show that the proposed method is fast and reliable for 3D reconstruction which has higher accuracy.