The Taub Faculty of Computer Science Events and Talks
Hadas Kogan (Elbit Aerospace Division)
Thursday, 25.05.2017, 11:30
Elbit Aerospace utilizes 3D mapping in a variety of different projects. We present two of these projects; in the first project, a 3D model of a large area is generated from aerial imaging. Here, the rapidly changing viewpoint of an aircraft mounted camera enables multi view geometry techniques to obtain a 3D point cloud. A mesh is reconstructed, and images from the same camera are used to create a photorealistic model. The composed model is 2.5D, that is, it includes one height for every lateral position. In our settings, 2.5 modeling is preferred over full 3D modeling, but it also introduces visual artifacts, for instance around bridges and trees. In addition, planar surfaces such as walls and roofs should be rectified in order to achieve a visual appealing model. We introduce a novel approach for improving the visual appearance and accuracy of the model. Our approach, for which we coined the term 2.75D modeling, combines 2.5 modeling for most of the area, and full 3D modeling only where necessary.
In the second project, a 3D model is generated during the flight and is used to compensate for degraded visual environment, e.g., during landing. In this case, a 3D model is obtained by combining a 3D point cloud scanned by a LiDAR, together with images from a camera to obtain a photorealistic model. The LiDAR provides a sparse point cloud that is accumulated while the craft maneuvers, using the craft’s position and orientation measured by a GPS\INS system. Accurate accumulation is crucial for good and reliable modeling. In cases where the position and orientation are not accurate enough, point cloud registration is essential. We present a novel approach for point cloud registration of sparse LiDAR signals.
Hadas Kogan leads the 3D vision team at Elbit Systems Aerospace division. She holds a B.Sc and an M.Sc from the Department of Electrical Engineering at the Technion, Israel Institute of Technology. Hadas has more than 10 years experience in conducting and supervising research and development at Elbit, HP-Labs and Rafael, in the fields of 3D modeling, computer vision, and machine learning.