Collecting my thoughts--
This week I opened up the two Artec scanners, the higher end Space Spider and the more mid-range Eva M. Both are older models that have been in storage. These are handheld scanners with laser/camera hardware.
I left the NextEngine scanner in storage for now because it is a desktop model (locked down instead of handheld) and I am familiar with it. Also, look at the NextEngine website.
It took some trial and error to begin getting results. Any scanning method has general requirements for the object to stitch properly. The process does not like shiny surfaces, repetitive patterns (both form/texture), and does not like objects to be too large or too small compared to the capture space. Some objects need extra treatment such as chalk dust, matte spray varnish, or capture enhancements such as strobe lights. And, some objects are simply better hand modeled. This French team has a fantastic exercise "Bien/Pas bien!" to help identify good objects for scanning.
With this in mind. I set up a table with the existing lazy susan (circular swivel platform?), a light, and collected a couple sizes of test objects. My setup is still a bit clunky. At one point I tripped over a scanner wire.
For one, the depth of field is very particular. Too close / too far and the live view disappears. The live view of the scanner is nice, but the software does not show the full image, instead it previews bits of the object that it is already trying to map. Because of this, I found medium-large sized objects to be easier to work with (relative to the scanner itself).
The Eva M scanner has been the most successful so far. Despite this older model having a lower resolution and a loud fan, the scan tests have been pretty decent. The Space Spider is more accurate and can get closer range, however he takes 15 min to warm up and seemed to be a little more finicky, I have yet to get a successful scan.
This week the software licenses will come in and I will actually be able to properly export. In the meantime I've been posting photos/video to my Instagram, @tasha.3d
I left the NextEngine scanner in storage for now because it is a desktop model (locked down instead of handheld) and I am familiar with it. Also, look at the NextEngine website.
It took some trial and error to begin getting results. Any scanning method has general requirements for the object to stitch properly. The process does not like shiny surfaces, repetitive patterns (both form/texture), and does not like objects to be too large or too small compared to the capture space. Some objects need extra treatment such as chalk dust, matte spray varnish, or capture enhancements such as strobe lights. And, some objects are simply better hand modeled. This French team has a fantastic exercise "Bien/Pas bien!" to help identify good objects for scanning.
With this in mind. I set up a table with the existing lazy susan (circular swivel platform?), a light, and collected a couple sizes of test objects. My setup is still a bit clunky. At one point I tripped over a scanner wire.
For one, the depth of field is very particular. Too close / too far and the live view disappears. The live view of the scanner is nice, but the software does not show the full image, instead it previews bits of the object that it is already trying to map. Because of this, I found medium-large sized objects to be easier to work with (relative to the scanner itself).
The Eva M scanner has been the most successful so far. Despite this older model having a lower resolution and a loud fan, the scan tests have been pretty decent. The Space Spider is more accurate and can get closer range, however he takes 15 min to warm up and seemed to be a little more finicky, I have yet to get a successful scan.
This week the software licenses will come in and I will actually be able to properly export. In the meantime I've been posting photos/video to my Instagram, @tasha.3d
No comments:
Post a Comment