Separating the bodies in the saved mesh

[24c]

I'm doing test scans to try and get to grips with the different lighting options on some vapour blasted parts. Sorry Dave aka Nerv still using the turntable and small parts whilst I get to understand the process. I'm saving .stl files and they are massive, a 90mm piston is around 720 MB, if it manages to save it without crashing (the software gets to the end of the "complete" operation and then crashes and reverts back to the desktop)and that's 3 passes or scans. One thing is the scans look great when they are on screen, but when you "complete" the process they get distorted with extra artifacts. I'm generally seeing flat areas become layered with scale like deposits if I do additional scans to see hidden areas.
One of the issues I did have was the parts were moving or rocking slightly when using the turntable as the movements are quite quick. Obviously didn't have that issue with a lazy susan icing board.
A couple of questions.
Is there a way of seeing or reverting these scans or bodies as separate entities?
Do you have to use the .asc files, that are saved in the format xxx_ScanData_n.asc and possibly layer & align in Meshlab for example?
If you don't like an intermediate scan can you delete it and "continue" again.

[Hugues]

I'm doing test scans to try and get to grips with the different lighting options on some vapour blasted parts. Sorry Dave aka Nerv still using the turntable and small parts whilst I get to understand the process. I'm saving .stl files and they are massive, a 90mm piston is around 720 MB, if it manages to save it without crashing (the software gets to the end of the "complete" operation and then crashes and reverts back to the desktop)and that's 3 passes or scans. One thing is the scans look great when they are on screen, but when you "complete" the process they get distorted with extra artifacts. I'm generally seeing flat areas become layered with scale like deposits if I do additional scans to see hidden areas.
One of the issues I did have was the parts were moving or rocking slightly when using the turntable as the movements are quite quick. Obviously didn't have that issue with a lazy susan icing board.
A couple of questions.
Is there a way of seeing or reverting these scans or bodies as separate entities?
Do you have to use the .asc files, that are saved in the format xxx_ScanData_n.asc and possibly layer & align in Meshlab for example?
If you don't like an intermediate scan can you delete it and "continue" again.

Hi Mike,

let me try to answer some of your points above, and i guess Dave and Alfred will add some more

- be sure you have enough disk space on your bootcamp partition, mine was too small and the software was crashing when running out of disk space. Probably a point to fix for Alfred.
- Concerning file size: did you select high level of details ? probably. Try medium or even low level, that will reduce the stl file size and you might have still enough details
- the extra artifacts come from meshing. WHen you see your scan on screen, they are not "water tight". Meshing tries to close your scan to produce an stl file. If you don't have 360 degrees scans, the system will try to close the back side any way it can and creates these artifacts.
- Be sure all your scans are aligned by the software, or else you might see these scales, maybe it's your problem, not sure. Post a picture.
- YOu can delete a scan you don't want but before you need to click on ALign (for 3 points alignment). There there is a delete button, down left. It brings you back to your previous scan. ALfred: would be nice to have the delete button too on the screen where we see the mis-alignment.
- the only way to see the separate scans is to save them as point clouds. I was also trying to align them using Meshlab, but it's a pain in the neck, Einscan software is much better at auto-aligning the scans than yourself in Meshlab. Most of my scans auto -align properly, you will get there too, it's just a matter of a couple of hours of practice.

Re-read the post of DAve on tips and advices, all is there. YOu need to overlap your scans. I believe scanning with tripod is easier than with turntable, you should scan without turntable as per Dave's advice (and now mine )

[24c]

Hi Hugues, thanks for the response.
I've seen the delete option in free scan, but after a few more hours I will be OK I reckon. I didn't find Meshlab too awkward to align, it's got a "fudge" factor like Einscan that gets you close. The only thing I did notice was the individual scans had artefacts outside the scanned object (turntable ring textures etc), but that'll disappear when I stop using the turntable. I'll try a free scan of the small parts as a double check. I've just been looking at light levels, and comparing the different res modes...and that was before I read your post. I did find the lowest res was picking up almost the same detail (letters in the casting)
My next step is try and convert the STL to STEP or similar. Trying Rhino, but not getting anywhere fast with this, and hopefully the CAD guys will update my Pro option that offers some mesh conversion tools.
I will re-read Dave's posts again.
Mike

[Hugues]

Hi Hugues, thanks for the response.
..The only thing I did notice was the individual scans had artefacts outside the scanned object (turntable ring textures etc), but that'll disappear when I stop using the turntable. ....
Mike

Oh, maybe your artefacts are coming from something else. Sometimes the scanner picks up unwanted objects in the background. You can delete them at the end of the scan, before aligning with the rest. It's written on the side, i think you need to press and hold shift and circled the unwanted scanned parts, they'll turn red, then you can delete them.

[24c]

Hi again Hugues, yep seen that... Fn-Delete on the MacBook to clear the red shapes.
I'm just trying a free scan at the moment, but for small objects it's more time consuming.
The other niggly thing, is my alignment keeps drifting, and although I select three point, the resultant model ignores the moves sometime, and I have to start the scan again as it's corrupted. Im on my fourth attempt, and can't get past five consecutive passes/scans so far...grrr. Last time I saved each, I'm[ported into Meshlab and used their alignment, that's how I knew it wasn't so bad. However, all this stuff is a learning curve.

[24c]

Alfred_t_s, how about you guys adding some drop downs from the menu in future, possibly even an edit tab, which might include an undo & redo function etc. Just another 2¢

[Hugues]

Alfred_t_s, how about you guys adding some drop downs from the menu in future, possibly even an edit tab, which might include an undo & redo function etc. Just another 2¢

+1 to that, the software is a little crude for sure.

[24c]

Anybody? How do you un-select if you've slipped with the green line, and caught your model... I can't seem to undelete. I only have a "Next" option. It got to be something simple.
Well we're at it, this model navigation when you're trying to align a small part is pretty poor. Are there any key commands for "fitting to screen" Having zoom zoom ability, tabbing to make a window active would be pretty useful too.
I couldn't work out how to un delete the artefacts outside the model, and now my main model window is very difficult to rotate and orientate the model. Someday's you just feel like you've got winter gloves on and cotton wool between your ears.

[24c]

Sorry guys, I twigged it, leave all the mistakes in place, go to the "Next" screen and "Delete" the model... and start again!

[24c]

Hugues, I tried all sorts of key combinations, to try and find some hidden ones!
It might be the MacBook Bootcamp keyboard mappings though.
Anyway, it's progressing better now...if I get a silly alignment, I just delete the model rotate the model back (equivalent of moving the camera too) and rescan... seems to work and a lot easier manual aligning
Don't forget these are small parts, Ø50mm and 120mm high.

PS Another quirk, when you're doing a scan and it aligns the to the model, the "cyan/green" area has to rotated for you to check...you'd think the scan model window would update and show you this. I rotate it to check the alignment and when it scan s again, it's at the back of the model again.... changes needed here Alfred_t_s... I know what it doing, it's referencing the start of the scan, not the active part.