Accurate and robust 3D scene reconstruction in the presence of misassociated features for aerial situational awareness

摘要：

　　Georeferencing based on aerial imagery has many applications,including remote sensing,environmental monitoring,map generation,and autonomous navigation of Unmanned Aerial Vehicles(UAV).In aerial imagery,Structure from Motion(SfM)is often used for ground point 3D reconstruction and camera pose estimation using geometrically matched features between 2D images.During SfM,misassociated features(i.e.,outliers)are excluded using an outlier rejection algorithm such as RANdom SAmple Consensus(RANSAC),but the outlier rejection algorithms cannot guarantee to return perfect feature correspondences.There might be potential misassociated features that are not identified.Consequently,those potential misassociated features will significantly lower the accuracy of 3D reconstruction.

关键词： bundle adjustment aerial sensing 3D reconstruction Georeferencing

作者: Mohammad Reza Jahanshahi Fu-Chen Chen Adnan Ansar Curtis W.Padgett DanielClouse David S.Bayard

作者单位: Civil Engineering,Purdue University,West Lafayette,USA Electrical and Computer Engineering,Purdue University,West Lafayette,USA NASA Jet Propulsion Laboratory,California Institute of Technology,Pasadena,USA

会议类型: 国际会议

会议名称: The 7th World Conference on Structural Control and Monitoring(7WCSCM)(第七届结构控制与监测世界大会)

会议地点: 青岛

会议语种:英文

页码: 1544-1545

在线出版日期: 2018-07-22（万方平台首次上网日期，不代表论文的发表时间）