The Skinny on Laser Scanning
by John Stebbins
About 10 years ago I began to realize that the whole world needs to be modeled. There is becoming a huge need for what are often called “digital as-builts” – 3D models based on how buildings are actually built.
Collecting 2D drawings to use as a basis for 3D BIM models of existing buildings could be a tricky affair, since there is no guarantee that the drawings have been truly updated and that all the building renovations have been included in the drawings on record. And what if the drawings are lost or the old blueprints are ragged and tattered or limited in scope? Have you ever tried to measure the height of the front facade of a football stadium with a steel carpenters tape or via triangulating using a builders level and trigonometry?
Enter laser scanning.
For those of you who have seen a laser level or laser surveyor’s transit, a laser scanner uses similar technology, sending out a laser beam that can bounce off surfaces and communicate back a distance that can be translated in software to x, y, z points. The file that results in a collection of x, y, z coordinate points in space is called a “point cloud”. These 3 dimensional point clouds can be quite dense. It is not uncommon to see single scan file sizes of 80MB or larger. Multiple scans are then registered or “stitched” together (rather like stitching photographic images together for panoramic QuickTime VR or IPIX scenes) to create a single unified point cloud. Once the scans are registered and unified, file sizes can easily surpass one or more gigabytes.
Special software is used to extract 2D or 3D data from the point clouds. This software is fairly complex and very expensive. For non-technical users there are now applications that can translate raw point cloud data into imagery that can be viewed with a free viewer program almost like a “.pdf viewer” but for scan data. The viewer software is very intuitive and doesn’t require any special knowledge of how to work with point cloud data. These images act a lot like a QuickTime object VR movie, allowing the user to orbit/spin 360 degrees horizontally and up to 310 degrees over the top. The viewer also has tools to measure distances from one point in space to the other, like the distance from the top of grading to top of parapet or the radius of an arch 3 stories high. Now, rather than going to the site with a long measuring tape, we can rely on the point cloud data to accurately give us actual distances based on scanning the building, as-built.
Think of the point cloud as a 3D connect-the-dots file, somewhat like a “vertices only” model file. What’s lacking is that these points are not triangulated into a surface model. Until someone comes up with 3D software that can interpolate the points and create a good 3D geometry-based model (still no intelligence), at least we have a file that can be “traced” with 3D intelligent BIM tools, like ArchiCAD, much like BIM modeling service bureaus trace 2D DWGs to create a 3D intelligent BIM model (2D to 3D conversion).
An API working inside of ArchiCAD would be a great solution for interpolating 3D point cloud data that can be translated into intelligent walls, windows, doors, etc. As cool as this kind of software could be, there is nothing like a knowledgeable 3D BIM modeler who understands buildings and knows how to interpret the point cloud to accurately “trace” the points and transform them into smart building elements.
The General Services Administration considers 3D laser scanning as one of the more mature BIM tools. Why would the GSA, the largest owner/operator in the world, want to create digital intelligent BIM models to improve the management and delivery of capital projects? First, it has value in the planning phase: it lets GSA provide architects and engineers with high-confidence data about existing conditions, a big improvement over handing them some archived drawings and saying, “Now go field verify it”. Next is its value during construction: to integrate the 3D model to the physical conditions of the site. Laser scanning technology insures that all building elements are aligned and provides verification that all as-builts are as-designed.
Digital as-builts are not only valuable during major remodels, but also for facilities management, and for a digital version of the ACTUAL spaces that can be used for energy analysis and management of the building’s entire life cycle.
As “the government’s landlord,” the GSA manages federal assets valued at nearly $500 billion. These assets include 8,900 government owned or leased buildings totaling 342 million square feet and housing 1.1 million federal workers in 2,200 communities nationwide. That’s a lot of world to be modeled!