More Than One Way to Scan a Monkey

1/16/2015

Getting a good scan and 3D print of smaller objects can be difficult without high-end equipment. As we shall see, it can also be challenging with sophisticated gear, as well. In a previous post, I discussed the ins and outs of scanning a 3d printing a small object, in this case a 8″ figurine of a monkey reading a book. Previously, I had scanned the Readin’ Monkey with the 3D Systems Sense handheld scanner, and created a 3D model with photographs and Autodesk’s 123D Catch software. The results were good, but not great, and I wanted to see if there was a better way to scan the Readin’ Monkey and get a good 3D print.

Matter and Form 3D Scanner

As far as consumer-level 3D scanners go, the 3DS Sense is good for medium- to large-sized objects (it gets great scans of people), but does rather poorly with smaller objects and details. Desiring the ability to get good scans of smaller objects, and hopefully greater detail than the Sense, I purchased the Matter and Form 3D scanner. This is a turntable scanner similar to the Makerbot Digitizer scanner. I opted for the Matter and Form scanner because it was: 1) less expensive, and 2) had a larger scan volume. I’m a sucker for a larger scan/build volume, something that has caused problems before.

3d scanning, 3d printing, tesseract, portland

The Readin’ Monkey on the Matter and Form scanner

The Readin’ Monkey was placed on the scanner’s turntable, which is several inches from the scan head. The turntable makes a 360-degree rotation, then the scan head moves up an inch or so and the process is repeated until the entire object is scanned. There are three big differences between the Matter and Form scanner and the Sense scanner:

  • The first is that with the Matter and Form scanner tracks automatically, so there is no chance of the scanner “losing track” of the scan.
  • The second is that the Matter and Form scanner uses a visible red laser (unlike the Sense’s IR laser), which is more sensitive to ambient lighting conditions and reflective glare from the object being scanned. In order to overcome the latter, one can lightly dust the scanned object with baby powder to give it a matte finish. (I chose to do this during rush hour, stepping out the front door of the building, which is on a busy street in Portland. I got plenty of bemused looks from passersby who were surprised by the sight of a man shaking baby powder on a monkey figurine on the side of Hawthorne Boulevard.)
  • The third is that the Matter and Form scanner “sees” the scan object linearly. In order to get a good overall 3D model of the Readin’ Monkey, multiple scans had to be taken, with the RM in different orientations. The Matter and Form software does a really good job of stitching together the scans into an overall digital model of the object, but as each scan can take upwards of an hour, and I took 7 scans of the Readin’ Monkey to ensure a good model, the process was time-consuming. Also, it can also take nearly an hour for the software to combine the individual scans into a unified whole, so the Matter and Form scanner can take the better part of a day to produce what the Sense scanner can do in a matter of minutes.

However, the proof of the pudding is in the quality of the scans. The Matter and Form scanner did, at the end of the day, produce a better scan than the Sense scanner. The following image shows a comparison on the two scans:

3d scanning, 3d printing, tesseract, portland

Comparison of scans from the Sense and Matter and Form scanners

Overall, the Matter and Form scanner performed pretty well. There were just a few drawbacks: the amount of time required to perform multiple scans, an irritating tendency for the scanner to lose connection with the computer, and some horizontal striation in the model. This is somewhat evident in the comparison image above, but became readily apparent when the model was printed on the Makerbot Replicator 2X.

3d scanning, 3d printing, tesseract, portland

Striations on 3D print from the Matter and Form scanner

At first, I was concerned that the striation was indicative of a problem with the printer. I’ve been tinkering around with the printer due to some print quality issues, so the striation issue may be a combination of a scanner issue and a printer issue. I’ll update as these issues get worked out.

Faro Focus Scanner

It just so happened that about this time I got my mitts on a high-end Faro Focus scanner, which I was using to scan the interior of a high school for another project. This is a heavy-duty piece of scanning equipment, which is used for taking 3D scans of building interiors and exteriors. (More details on this in an upcoming post, but you can see and example of a building scan here.) I took the opportunity to use this scanner to scan the Readin’ Monkey, thinking that I would be able to get a really high-quality scan. While I did get a good scan, I had problems processing the output of the scanner into something that would be usable by a 3D printer.

Without going into too much detail, the output of the Faro Focus is a raw point cloud of X,Y and Z coordinates, whereas the output of the Sense and Matter and Form scanners are printer-ready STL files. The problem that I encountered was that the Focus shoots so many laser points that there is a fair amount of scattering that occurs along the edges of the target. I took eight scans of the Readin’ Monkey with the Focus, and when they were stitched all together, there was a wonderfully detailed scan of the RM, but it was surrounded by a nimbus of random scattered points.

3d scanning, 3d printing, tesseract, portland

Faro Focus Readin’ Monkey scan with scattered scan points

When converted directly to an STL file, the result is a “blobby mess” that I didn’t even try to print. The trick is in finding a way to filter out points that are a certain distance away from other points. The software the accompanies the Focus is incredibly difficult to use, and I was not able to figure out how to successfully filter out the unwanted points. I suspect that other apps such as Blender or MeshLab will be able to accomplish this, and I’m looking forward to seeing what can be achieved with these tools.

Overall, it’s been very interesting to see how the Readin’ Monkey can be scanned and replicated with a 3D printer. So far, the Matter and Form scanner has produced the best scanning results. (I’m still working on ways to eliminate the striation issue.) As I run across other ways to scan and print the RM, I will be sure to post the results to this blog.

—-

For more information on how Tesseract Design can help you with your 3D printing and scanning projects, please see www.tesseract-design.com/3-d-printing