Video: BLK2Go, aurivus AI, and Revit

In the video below we demonstrate how our AI performs on a scan from hand held laser scanners. In this case we used demo data from the Leica BLK2Go.

Usually, the point cloud from a laser scanner is made from one single piece. It can be hard to model objects when the modeler has to look through a wall made from points. Thanks to our neural network (AI), millions of points are grouped into only a few objects. Thus the modeler can select parts or classes of interest.

Our AI is trained to “understand” point clouds from building scans. The task for which we trained our AI is to divide the raw point cloud into single groups, e.g., single windows, and tag every group, e.g., as Window. The result is a “smart” point cloud that can be loaded into CAD software, e.g., Autodesk Revit.

Comments are closed.