In “LayerCode: Optical Barcodes for 3D Printed Shapes,” researchers Henrique Teles Maia, Dingzeu Li, Yuan Yang, and Changxi Zheng have created a tagging scheme called LayerCode. This embedding process tags 3D printed objects with bar codes during production, allowing them to transfer data about parts and objects without affecting the form of the print at all.
LayerCode tags are compatible with FDM 3D printers and SLA-based printers, whether in producing complex or ‘nontrivial’ shapes. This process is an improvement over those in the past, according to the researchers, which may have failed for numerous reasons, but normally due to issues with placement. In this study though, LayerCode was successful 99 percent of the time, with 4,835 shapes tested during this study.
As the authors of the research project point out, barcodes are considered by many retailers and manufacturers to be a necessity, but their intrinsic qualities with standard black bars are also somewhat like the structure of layering in 3D printing. In terms of attaching them to the 3D prints, however, that has historically been a difficult process due to the varied, customized geometries of so many objects. LayerCode was developed exactly for those purposes, embedding codes in curved or rounded parts.
As the project began, the research team’s goal was to embed data in the printing layers with ‘robust encoding,’ along with a specific algorithm for encoding information about the object or component. They realized the need for ‘two distinguishable layer types,’ creating a new coding mechanism, along with software and hardware updates for the 3D printers in use.
For printers using up to two materials, layer types were introduced with varying materials. In FDM single material printers, they changed filament deposition height. For SLA printers, they mixed near infrared (NIR) dye into the resin to make a second layer type.
“Our proposed LayerCode approach features a number of attributes desired for tagging 3D printed objects,” stated the researchers.
These attributes include:
- Robustness on complex shapes
- Ease with a conventional camera
- Compatibility with 3D printers
- Structural and appearance preservation
- Object tagging
- Depth information
As they began printing, testing on all three types of 3D printers, the researchers noted that jobs took from 40 minutes to three hours, based on how complicated the object was.
“When printing with two colors and with variable layer heights, the time costs are comparable to printing without LayerCode tags. For Ember printing with NIR resin, we observed a minor overhead (about 15% to 20% slow-down) because of the extra tray swaps,” stated the researchers. “A remarkable strength of LayerCode tags is the ability to decode even when the object is damaged, thanks to the layer-by-layer printing process that spreads the tag over the entire body of the object.”
The authors also realized that LayerCode tags would be useful in steganography for 3D printing, with a watermark left on each item to deter counterfeiting.
They went on to test the decoding algorithm, using synthetic images from a photorealistic renderer. During the assessment, they realized it was not cost-effective to 3D print each shape from the datasets they created. With one object taking up to an hour to print, virtual rendering would be the better choice, but they stated that it was ‘desirable’ to test their algorithms on the shapes ‘to prepare for the future,’ as they expect greater affordability and speed in 3D printing soon.
With the virtual environment, however, they were able to many 3D printed items from a variety of different angles, giving valuable insight for further development of LayerCode tags. Testing revealed that out of 4,835, only 44 could not be decoded.
“If an object is completely occluded or poorly illuminated, decoding will fail. The ability to decode also depends on the camera view angle. While as shown in our experiments, LayerCode tags can be correctly read from a wide range of camera angles, there are other view angles (such as those nearly aligned with the printing direction) from which the decoding is prone to failure,” stated the researchers in their conclusion. “Therefore, optimizing for how a shape might be held, seen standing, or made less visible would certainly improve robustness.
“Similarly, since not all angles are equally easy to decode, processing multiple views in parallel to achieve more robust decoding also serves as an exciting avenue for future work.”
While so much of 3D printing innovation still emanates from the home workshops of makers, hackers, and tinkerers around the world, as it becomes more entrenched within important industries, we will see many interesting changes—as with embedding bar codes, other interesting identifiers and retail agents may be in place such as luxury goods authenticity tags or other inventory tracking. What do you think of this news? Let us know your thoughts! Join the discussion of this and other 3D printing topics at 3DPrintBoard.com.
[Source: LayerCode: Optical Barcodes for 3D Printed Shapes]