Demonstration of Urban Robotics' Dense 3D Extraction capability using aerial imagery of The Banfield in Portland, Oregon. Via a fully automated process, image calibration, optimization and 3D extraction was performed in a single action, using 30 source images.
.
This video renders a dense 3D colorized point cloud extracted from the source 30 images. The resultant point cloud consists of over 160 million points. Each point is a pixel from an image that has been reprojected into its position in 3D virtual space. The process is entirely automated and requires no external data sources. In particular, no lidar data was used to create this data - just 35mm cameras.
.
If you are an engineer/scientist skilled in the art of multi-view geometries, 3D reconstruction and general computer vision, please note that we are actively hiring experts in the field. Please contact us through the email address at the end of the video.
Link to this comment:
All Comments (0)