Capturing Reality: Architectural 3D Scanning and Modeling

buy Lyrica online india 1/31/2015

One of the big challenges of an architectural renovation project is obtaining accurate documentation of existing conditions. As anyone who has worked on such a project can tell you, “as built” drawings can be anything but. Even for relatively recent projects, the task of accurately documenting the final built project often gets overlooked, or done in a hurry. For older or larger buildings, previous renovations may not have been documented at all, or those drawings have been lost. Relying on existing documentation can be a risky bet. Sometimes these drawings deviate from real conditions by a significant amount. If not noticed in time, these deviations can result in re-design delays or costly changes in the field.

Fortunately, 3D scanning offers a thorough way to accurately document existing conditions for buildings and spaces, regardless of size and complexity.

website link 3D Laser Scanning vs. Tape Measure & Clipboard

3D laser scanning is an unbeatable way to get fast, thorough and highly accurate documentation of existing building conditions. In this blog’s very first post, I briefly touched on this aspect of 3D scanning. I recently had a chance to do a 3D scanning and modeling project, and would like to cover this in more detail.

The traditional way to document existing conditions was to send a couple of interns to the site with a tape measure and clipboard to measure and record the important dimensions of the space. There are a number of drawbacks to this method: it’s time-consuming, open to potential inaccuracies, and not very thorough. With 3D laser scanning, however, one can quickly develop an accurate representation of the space that can be very easily converted to a 3D model in Revit.

The scans are also treasure troves of information, and it is a simple matter to directly open a scan and pull a measurement, something that is extremely limited with the traditional tape-measure-and-clipboard method. Need to know the distance from the top screw of the lightswitch plate to the end of the third slat of the Venetian blind? Easy to reference with a 3D scan.

3d scanning, bim, tesseract, portland

Getting post-facto measurements from a 3D scan is easy.

Some may be concerned that the cost of 3D scanning may be prohibitively expensive, but this is just not so. A large construction company recently compared two jobs of similar scope where they needed to document existing conditions. The first, a building in downtown Portland, was done with a crew using a hand-held laser measuring tool (a slight step up from a tape measure). This was done with a crew that cost $75/man-hour and took 800 man-hours, for a total cost of $60,000. The other project, a building on the University of Washington campus, was done with a 3D scanning crew costing $150/man-hour. Despite the more expensive outlay, the scanning was completed using 240 man-hours of time, for a total cost of $36,000. So, the construction firm not only spent less time and money, they came away with a much more accurate record of the existing conditions.

Anatomy of the 3D Scanning Process

Recently, Tesseract Design was commissioned to scan and model portions of a high school outside of Eugene, Oregon. We used a Faro Focus scanner to do the job. The function of the Focus was described in detail in a previous post, but to summarize, the Focus shoots out hundreds of thousands of laser pulses per second, returning a point cloud of data of the surrounding environment.

Each point cloud represents a line-of-sight view of what the scanner can “see” at any one position. There is an art and science to planning the series of scans in order to get the most information with the fewest number of scans. Looking at the existing drawings can be useful in developing an overall scan strategy, but as I’ve already mentioned, relying on as-builts is a dubious endeavor. One of the biggest problems is that there is often furnishings in the spaces that can block out large parts of the scans. In the case of the school in Eugene, I encountered an issue right away: the locker room was full of lockers! (Who’d have figured?) These lockers were six feet tall, and of course not indicated on the drawings, so I had to wing it.

While planning is important to 3D scanning, even more vital is to have targets. Lots and lots of targets. My rule of thumb is to come up with the number of targets that you think would be reasonable, and then double that number. In a scanning project, the scanning part is easy. The tricky part is registration, which is the process by which the individual scans are knitted together to form a unified whole. Having a cartload of targets in the scans makes this difficult job somewhat easier. Or, in some cases, prevents a difficult job from being downright impossible.

There are two types of targets that can be used: spheres and checkerboards. The spheres are handy in that they can be seen from a number of different angles, but they can be hard to identify in the registration software. I tend to rely on printed checkerboard targets, which I print with an identification number so I can be sure to keep track of which target is which.

3d scnning, bim, revit, tesseract, portland

Scanner’s-eye view of a 3D scan. Green circle indicates a sphere target, checkerboard targets have black-and-yellow crosses. The “lunchbox” objects represent the placement of the scanner for other scans.

 Registering the 3D Scans

Once the scanning is done, the challenge is to register the scans to a unified whole. The Faro Focus scanner is very easy to use; the Scene software used to register the Faro scans is not. In fact, it is one of the most ornery, non-intuitive software packages I’ve ever had to use. Other applications, such as Autodesk’s ReCap promise an easier time of scan registration, but for this project I opted to deal with the devil I knew rather than grapple with an unknown.

Fortunately, the process went pretty smoothly. I had enough targets scattered about so that Scene was able to register a large number of the scans automatically. The only hitch was the transition where I went from indoors to outdoors, which Scene found confusing. I had to “force” a number of targets – that is, manually require the software to match targets by name. I also had to use some of the features of the scan (flat walls or planes) to help Scene make some fine-tuned alignments. Overall, the registered scans came out pretty well, although there were times when Scene severely tried my patience.

3d scanning, bim, tesseract, portland

Registered scan of the locker room area, birds-eye view.

Pulling it into Revit

The final step in the project was pulling the scan into Revit and modeling the scanned features. The biggest obstacle was a hitch in which the Faro software and Revit communicate with each other. This is a known issue, but both Faro and Autodesk seem to expect the other to tackle the solution. In the end, I found a workaround by exporting the registered pointcloud in a LIDAR format, and importing that into Revit. Once in Revit, I could simply begin modeling over the 3D pointcloud to develop an accurate model of the high school. There are tools and workflows that help automate this process, but in this case I preferred to stick with simply using the 3D pointcloud as the basis of the Revit model. The results were quite satisfactory, and there were features of the finished Revit model that would have been difficult to accurately model with the old-fashioned tape-measure method.

3d scanning, Revit, bim, tesseract, portland

Exterior of gym with locker interiors modeled over the pointcloud in Revit


For more information on how Tesseract Design can help you harness the power of 3D scanning, see