CAMheads CNC Router Forum By: CAMaster CNC

Go Back   CAMheads CNC Router Forum By: CAMaster CNC > Tips & Techniques > 3d Scanning & Equipment

Reply
 
Thread Tools Display Modes
  #11  
Old 08-17-2019, 03:09 PM
Terry Williams Terry Williams is offline
Cobra Owner
 
Join Date: Mar 2012
Location: Naples, FL
Posts: 239
Default

Awesome!

How long did it take to get it done, did you do it with a handheld scanner?

I have scanned some large objects with my scanner in handheld mode, but it takes a while - especially if it's hi-resolution. I have to use a powder spray (like Spotcheck) and markers to get a scan result that good on dark objects like oiled brass.

Great Job!

This is a pretty cool project - I'm an an old Machinist Mate (USS Enterprise CVAN-65 1969-1975) so I can appreciate it.

Thanks!
__________________
Terry Williams
Cobra 510 with Recoil
Reply With Quote
  #12  
Old 08-22-2019, 08:34 AM
BradyWatson BradyWatson is offline
CAMhead Legend
 
Join Date: Oct 2014
Location: 1 hour South of Philly
Posts: 339
Default

Terry,
Thank you for your service & your kind words.

No, this was not done with a handheld scanner for various reasons - but mostly because handhelds (for the most part) would require hundreds of registration markers all over the model to match up overlapping scans and to track the handheld in 3D space...The other thing about handhelds is that it takes an incredible amount of hand/eye coordination in order to keep the scanner in the 'sweet spot' where measuring is most accurate and linear - using in the center of the depth of field range. If you don't maintain the proper working distance from the model being scanned you will either have problems matching up adjacent scans or the data just won't be good...the list goes on.

The only prep to the bell I did was to add a bit of talc to the black shiny lettering to reduce reflectivity and to bring up the signal strength in comparison to the surrounding areas.

In this case, since the bell was outside, I brought a 12x12' black enclosure with me to block out the sun and give me enough working space around the bell. I had to move it a few times. It did a pretty good job of blocking the sun, but there was still more light than I anticipated so I had to do some fancy footwork to get the results I wanted. As you can imagine, it got wicked hot in there...so I had to pop out and cool off from time to time. It took me about 2hrs to setup, scan and break everything down after I verified I had full coverage of the bell.

Incidentally, someone else had tried to 'scan' the bell using photogrammetry - where a series of photos are taken (100+) and then stitched together to create a 3D model. The problem is, that doesn't work so well on models that are large with fine detail, such as the lettering and date on the bell, which had a hand-punched texture and max depth of only about 1/8" or so. Furthermore, the lettering was all done by hand back in 1905/1906 so it isn't like you are going to match the font. Even if you unwrap the bell and attempt to create them from a photo (digitize/trace then subtract in Aspire etc) you are up against compound rotary or spherical distortion, that bends things in ways you really can't account for in software...Studying that sort of thing is an interesting 'case study' for those interested in rotary type work and the effects of rotary/spherical distortion.

Once I had the data back in the office, I was able to analyse it and clean it up. The raw scan data was somewhere in the 2GB range, so that had to be decimated to a more reasonable size on disk so that it could be used for 3D printing. Rather than use the data as-is for the top nut/yoke where the bell is suspended from, I rebuilt those portions in CAD and added it to the rest of the bell. The original parts were rusted and had paint/rust flaking off so they were pretty ugly in real life.

It was interesting for me to see just how far off my measurements were using photos and creating the bell from cross sections etc compared to the 'true to form' measurements from the scanner. Even though I've been doing this type of work for 2 decades, every project has an opportunity to learn something new or to try out new things.

-B
Attached Images
File Type: jpg 3DBellWithVectors.jpg (61.3 KB, 13 views)
__________________
IBILD Solutions - High Definition 3D Laser Scanning Services - Advanced CNC Training and Consultation - Vectric Custom Video Training IBILD.com
Reply With Quote
  #13  
Old 08-23-2019, 10:19 PM
Terry Williams Terry Williams is offline
Cobra Owner
 
Join Date: Mar 2012
Location: Naples, FL
Posts: 239
Default

Yep, I figured this was no walk in the park.

Thanks for taking the time to explain the process.

You are right about those pesky markers - I have to use them for hand held scans and have to keep around 5 or so in frame at a time to keep from losing registration. If the object is small enough to put on a turntable, I can usually get pretty decent results without them, but it takes some setup time.

I use spotcheck for dark and shiny objects - it seems to work pretty good, but it's not cheap and you have to clean it up after the fact.

There is a new product from Europe called AESUB Blue Vanishing 3D Scanning Spray that solves the clean-up issue. It It is a relatively inert spray powder that disappears in a few hours with no cleanup (sublimation). I'ts a bit pricey though at $45/13oz can with $25 shipping.

Post more when you can, it's pretty interesting stuff!
__________________
Terry Williams
Cobra 510 with Recoil
Reply With Quote
  #14  
Old 08-23-2019, 10:37 PM
Terry Williams Terry Williams is offline
Cobra Owner
 
Join Date: Mar 2012
Location: Naples, FL
Posts: 239
Default

Here's a handheld scan of a conch shell - about 8" x 4" without markers or powder - just happened to have a surface that was scanner friendly.
Attached Images
File Type: jpg conchscan.jpg (20.0 KB, 15 views)
__________________
Terry Williams
Cobra 510 with Recoil
Reply With Quote
  #15  
Old 08-24-2019, 11:07 AM
BradyWatson BradyWatson is offline
CAMhead Legend
 
Join Date: Oct 2014
Location: 1 hour South of Philly
Posts: 339
Default

Terry,
I used developer exactly once - it is pretty nasty stuff and the scan lab is a cleanroom...so that doesn't work out so well when you have to manipulate/pose the part and you get it on everything. About 15 years ago I did my own research/case study on 'scan sprays' and just wound up developing my own that won't hurt the model being scanned and stays on when touched with your fingers, but comes off with a microfiber cloth. It pays to do your own R&D with things like this.

Nice shell scan. They can be trickier than one would think because some are semi-translucent. This causes the light source to penetrate the surface and report a deeper measurement than the actual surface - which results in noise. Looks like you lucked out with this one.

I scanned this African lion bust this morning...Believe it or not, it's hard to find one that looks 'right'. This one is close, but the nose is crooked and eyes are not exactly dead on, so I will probably do some hacking and sculpting on this one to get it where I want it to be.
Attached Images
File Type: jpg LionScan.jpg (99.8 KB, 27 views)
__________________
IBILD Solutions - High Definition 3D Laser Scanning Services - Advanced CNC Training and Consultation - Vectric Custom Video Training IBILD.com
Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -4. The time now is 02:45 AM.


Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.