Services,

GHVM has developed a suite of software GeoConstructor  that allows us to efficiently acquire the data and photographs necessary to construct large scale, high optical resolution models.  Models that contain >20 million triangles and >800 photographs (24.5 Mpx) have been assembled for kilometer scale geological outcrops.  The models were acquired using high speed terrestrial LiDAR at a range of 60 to 500m and a robotic camera system with a 300mm lens.  The optical resolution of the models was 1.5 mm/pixel in the areas of interest.

In order to create a photorealistic model of a geological outcrop, or man-made structure, the surface of the outcrop is captured using high speed terrestrial LiDAR scanners.  The point cloud image of the outcrop is converted into a TIN mesh surface.  Photographs are then draped onto the mesh to create a photorealistic model of the outcrop.  The draping is accomplished by mapping each vertex in the mesh to its appropriate point in a photograph.  The mapping process requires that each photograph be oriented to the model in exactly the same orientation as the camera to the outcrop at the time that the photograph was taken.  The orientation step requires at least four tie-points between the photograph and the model.  Usually, this is done manually resulting in one or two photographs being processed per hour and often with poor accuracy.  In order to efficiently drape a large number of photographs onto a large scale TIN mesh, it is necessary to dramatically increase the speed with which the photograph-to-model tie-points are identified.  Our GeoConstructor software combines a robotic total station, a robotic pan-tilt stage, and a large format (24.5 Mpx) professional DSLR camera with telescopic lenses to create an automated system for outcrop photograph acquisition and photograph tie-point registration.