Scan It, Snap It, Print It

Using 3D Scanning and Photogrammetry to Create Content for 3D Printing

9/7/2014

A Chicken in Every Pot and a 3D Printer in Every Home

Get More Info There’s no question that 3D printing is going to usher in some major changes to the way items are designed and built, from nano-sized widgets, all the way up to entire buildings. However, right now nobody seems to really understand what form those changes are going to take or how they are going to happen. The democratization of 3D printing has made this amazing technology available to consumers and hobbyists, with 3D printers now available at retail outlets such as Staples. Optimistic 3D printing enthusiasts predict that there will soon be a day when there is a 3D printer in almost every home in the U.S.

look what i found While this scenario may very well come to pass, one thing that isn’t certain is how content for this plethora of 3D printers is going to be created. The common analogy between 3D printing and 2D printing breaks down pretty quickly after a certain point. Most folks have a word processing application on their computers, and even if they never had any formal training, it was a relatively simple matter to hunt and peck out a document that could be printed on a desktop printer. With 3D printing, however, it’s a little more complicated. While 3D modeling programs are becoming more common, cheap, and easier to use, the ability to use these tools to create complex 3d models for printing can be quite time-consuming and difficult.

Granted, there are ever-increasing libraries of 3D printable content, such as Makerbot’s Thingiverse. However, the would-be 3D printer is still limited to something that someone else created. This potentially limits the market for the household 3D printer scenario.

3D Copying Reality

Regardless of the steep learning curve of 3D modelling tools, there are other ways for the would-be 3D printer to create content to print. Low-cost 3D scanners and photogrammetry tools are available that can capture real-life 3D objects and save them in a 3D file format. Easy-to-use software tools can be used that can clean up and edit these captures, and get them ready to print. I’ll take a look at two of these tools that I’ve used with varying levels of success: the 3D Systems Sense handheld scanner and Autodesk’s 123D Catch photogrammetry tool.

Sense 3D Scanner

3D Systems created major buzz when they announced the Sense scanner last fall. As mentioned in a previous post, they seemed to have taken competitors by surprise (especially Makerbot). The Sense has a great price point – nearly an order of magnitude less expensive than the next tier of handheld 3D scanners.Makerbot’s Digitizer comes close, but is still nearly twice the price of the Sense, and is limited in the size of objects it can scan. There are also some pretty interesting 3d scanner hacks that can be made from the MS Kinect system, but these require a certain level of DIY handiness, and often a large and cumbersome framework on which to mount the apparatus. Overall, the Sense is the best bang for the buck for a portable, easy-to-use 3D scanning system.

123D Catch

Autodesk’s 123D Catch is a photogrammetry application that takes a series of photographs and creates a set of points from which a 3D model is created. There have been a number of similar applications in recent years, but it is only relatively recently that these apps have really gotten the fine-tuning they’ve needed to work really well. 123D Catch requires about 40 photographs to create a workable model. The photographs are uploaded to the ever-popular “cloud” for processing on Autodesk’s servers. The finished result is then downloaded back to the user’s PC. This can be time-consuming, depending on connection speed, desired resolution and the load of work that might be going on at the server farm. On the other hand, as a free app, the price is unbeatable.

The Monkey Test

In order to test the relative merits of the Sense scanner at the 123D Catch application, I decided to do a side-by-side comparison. The model was an 8″ figurine of a monkey reading a book. The monkey was scanned and photographed, and the resulting 3D models printed on a Makerbot Replicator 2X printer at 0.2mm resolution.

3d printing, 3d scanning, tesseract desgin, portland

The “reading monkey” with the resulting printed models. On the right, the 123D Catch model, on the left, the Sense scanner model.

Blue Monkey – Sense Scanner

The Sense scanner is handheld, and looks like a staple gun. It attaches to the PC via a USB cable. The built-in 6′ cable isn’t nearly long enough, so I’ve added a 12′ extension, and that usually provides enough length to scan even large objects. At 8″, the monkey figurine definitely doesn’t fall into the “large” category for the Sense. The accompanying software has presets for the object size to be scanned, and under the “small” category are objects like laptop PC and basketball.

Since the Sense is a handheld scanner, the use must be careful not to move it too fast or it will lose the tracking of the object. The Sense software displays what the scanner is “seeing,” so you really need to be able to see the computer screen during the scanning process. Even when great care is taken, the Sense can sometimes lose the tracking of the object. It is possible to recover the tracking once it is lost, but only about 50% of the time. The other times, you basically just have to start over. This is the biggest drawback to using a handheld scanner, and it can make for some frustrating scanning. It took about four tries to get a good, complete scan of the monkey figurine.

Red Monkey – 123D Catch

The tracking frustrations inherent in using a handheld scanner are absent when using a photogrammetry application like 123D Catch. You just have to take lots and lots of pictures. It’s also very helpful if you tape down newspaper with a variety of colors and shapes underneath the target object. This helps the application knit together all of the photos into a 3D model. It is also helpful if you have to go back, and manually register some of the photos if the software can figure it out automatically. Fortunately, this does not happen very often.

In order to obtain the best results, you need to make two laps of the target object, taking around 20 photographs for each lap. The first pass should be slightly above the target object, looking down, while the second should be level with the centerline of the object. Once the photographs are done being processed, 123D Catch can show where the pictures were taken.

3d printing, 3d scanning, tesseract design, portland

Spiral path of photographs for the 123D Catch version of the monkey.

 And the Winner Is…

Well, I don’t know if it’s fair to say that there is a winner or loser in this competition. Both the Sense scanner and the 123D Catch app both produced passable models of the reading monkey. Neither were highly detailed, which was something of a disappointment, but not exceptionally surprising. The original figurine has a small pair of wire-frame glasses, but neither method picked up this level of detail. These are not super-detailed methods of 3D reality capture – that requires a considerably larger outlay of cash.

Overall, I’d have to give the edge to the Sense scanner. The model that it produced was by and large more true to form of the original. The 123D Catch model tended to be a little lumpy. This was particularly noticeable on the base of the figurine, which is a smooth and regular shape.

3d scanning, 3d printing, tesseract design, portland

123D Catch produced some noticeable irregularities in the base.

I liked the Sense scanner for the relatively high quality of the 3D model that it produced. Given the relatively small size of the object, it did very well. The performance of 123D Catch was also impressive, especially given that it is a free app/service. I deliberately chose a challenging object for this test, as I knew both methods of 3D capture have challenges at this scale.

3d scanning, 3D printing, tesseract design, portland

The Sense scanner model gets the nod for overall best 3D model.

I’ve done scans and photogrammetric captures of similar objects at different scales, and each of the methods have their strengths and weaknesses. For scanning people, for instance, the Sense wins hands-down. The 123D Catch models I’ve done of people resulted in the heads and faces looking very weird and distorted. However, the 123D Catch does manage to capture smaller items that the Sense just can’t handle. A smaller monkey figure (yes, I like monkeys) that was about 2″ x 1″ produced a recognizable model from 123D Catch; the Sense scan of the same object resulted in an elongated blob.

There is more than one way to scan a monkey, and in today’s post I presented two: using the Sense handheld scanner, and taking 40-plus pictures and stitching them into a 3D model using the 123D Catch application. Neither method was perfectly detailed, but overall they produced decent-looking 3D prints in ABS from a Makerbot Replicator 2.

As the use of 3D printers by hobbyists and tradesmen becomes more widespread, we can expect to see an increase in the quality and level-of-detail available for tools and apps like the ones I tried out here. Keep checking back to this blog, as I will let you know about them as soon as I hear about them. In the meantime, happy scanning!

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