Dear DAVID Community,

As you are probably aware, the entire DAVID company, team and product range has been integrated into HP Inc.
This forum is now read-only. Relevant content has been migrated to HP and merged into HP's Support Forums on November 1, 2016.

To start new discussions for 3D Scanning, please register and post your new topic at the HP Support Forums

Calibration Panel - left and right inner point

The place for questions, problems, comments and tips regarding the camera calibration.

Re: Calibration Panel - left and right inner point

Postby hunkatibor2 » Wed Jul 09, 2014 6:43 am

Hi
After reading the above in writing, performed the calibration result is.
Tibor
Attachments
2014-07-09_07-33-42.png
2014-07-09_07-35-43.png
2014-07-09_07-38-36.png
Intel Xeon E5410 2.33GHZ 16GB RAM Windows 7 64bit David 4.5.0
2 Basler A622F camera 2 Tamron Asfirical AF 28-80mm lense
LG DLP Projector PF1500G Canon D40 Tokina 17-70mm 1:2.8-4,5
Humanti Calibration panel system and Alignmed&Two camera
hunkatibor2
 
Posts: 1408
Joined: Tue Mar 16, 2010 4:26 pm
Location: Hungary

Re: Calibration Panel - left and right inner point

Postby sts » Wed Jul 09, 2014 12:03 pm

MagWeb wrote:That looks very good. Seeing this, I don't think there's a calibration problem and your partial pipe scans should be right too...

Think it's an aglinment problem where DAVID finds a "better" fit of the to regular surfaces if it sets another orientation than the right one.
Did you try to use ContactPair? - Maybe you could apply some 3D markers on the surface (some dots of clay or something)? That makes it easier to find corresponding matching points on the partial scans.


I tried normal aglinment, i tried ContactPair, i tried use texture. I now tryed dots (little mat dots sticker). Does not help.

But i think i have found the problem:
The midpoint of one of the "doughnut circles" is not a row of the midpoints from the six simple dots. Its 7 mm off its destination.

So i will build new panels. But than i have some questions, becase on the second try i wanna make everything right:
- The distance between two points on the SLS2 glas panel is 7.5 times the diameter of one point (Range: 240mm / Distance 60mm / Dot 8 mm). Is this value important or is roughly around this size okay? My dots are 8.6 times in one distance.
- My inner diameter of the "doughnut" is equal (like the SLS2 panel)to the normal dots but half the diameter of the outer "dougnut". The original iss smaller (Range: 240 / Dot 8 / doughnut inner diameter 8 / doughnut outer diameter 14). I think it is not important becase the range 120mm panel uses doughnut inner diameter 8 / doughnut outer diameter 14 too. So the doughnut is by far larger than the normal dots in the 120mm range panel.
- I use mat white resopal / formica compound plates. To get the the dots i use a cnc mill and cut out 2mm from the surface. Unter the surface, the plates are no more white, instead they are very dark brown. But can i get problems with the 2mm pit in the plate?

sts
Attachments
calib_panel.png
sts
 
Posts: 16
Joined: Fri Jun 27, 2014 1:35 pm

Re: Calibration Panel - left and right inner point

Postby MagWeb » Wed Jul 09, 2014 1:11 pm

You mean the one ring marker is displaced on the original boards shipping with the SLS-2 kit?
Could you post a shot?

A rough size of the makers and a rough ratio of ring and white dot is OK.

Engarving gives one problem: The visible shoulders of the engarving reflect the light different than the ground of the engarving. In certain light/contrast settings this might be a problem if DAVID decides to use the ground only... I use to fill the engarvings with black putty and sanding it carefully.
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Calibration Panel - left and right inner point

Postby sts » Wed Jul 09, 2014 1:41 pm

The problem was an mistake by me with the rebuild of the panel in the CAD program.

The SLS2 glas panel is fine (ive used the 120mm and 240mm for serveral objects). The PDF, CDR and the PNG that comes wird the David Software are great too.

Ok, then i will close the holes.

sts
sts
 
Posts: 16
Joined: Fri Jun 27, 2014 1:35 pm

Re: Calibration Panel - left and right inner point

Postby NARMELLA » Wed Jul 09, 2014 2:34 pm

Hi MagWeb

To check whether your results are distorted:
-Do a fresh calibration (best with all settings at default)
-Do not (re)move the corner but use it as an object to be scanned.
-Scan it
.
.
.


You're so smart.

It's a great idea to check the precision of calibration.

And it would be good to have a result of calib surface deviation analysis by david.(I don't know is it possible in future or not)

Thank you to your useful suggestion

NARMELLA
Data projector : NEC VE 281
Camera(s) : Canon eos 600D (3x)
Software : Lets scan 2.5 + David laser scanner 3.9
گشتاور صنعت تبریز
User avatar
NARMELLA
 
Posts: 354
Joined: Thu May 29, 2014 9:18 am

Re: Calibration Panel - left and right inner point

Postby hunkatibor2 » Wed Jul 09, 2014 4:08 pm

Hi Magweb
I have a question. The original embedded software in virtual box by default perfectly meets? My picture is a big gap. Send me a picture of the perfect virtual default calibrated and tested panel. Any mistake can be derived from.
Tibor
Intel Xeon E5410 2.33GHZ 16GB RAM Windows 7 64bit David 4.5.0
2 Basler A622F camera 2 Tamron Asfirical AF 28-80mm lense
LG DLP Projector PF1500G Canon D40 Tokina 17-70mm 1:2.8-4,5
Humanti Calibration panel system and Alignmed&Two camera
hunkatibor2
 
Posts: 1408
Joined: Tue Mar 16, 2010 4:26 pm
Location: Hungary

Re: Calibration Panel - left and right inner point

Postby Sven » Wed Jul 09, 2014 4:32 pm

Hi STS,

Gunter has already given great tipps and answered your questions, correctly of course, I just want to confirm officially:

- You don't need to change any Advanced Settings unless you build a completely different calibration corner.
- It does not matter if the red and green distances (your first post) have the same length. They don't. DAVID can compensate such offsets ("CompensatePatternOffsets"). It's just not possible to produce foldable perfect calibration corners at a reasonable price. Also when you use the back side (240mm) the offsets will be different anyway. That's why we made DAVID compensate that.
- When you build your own calibration corner, make sure the patterns have no error in ROTATION.
- "ExpectedPlaneOffsetsOfOuterCorner" is for "negative corners" only, not relevant here.
- When "CompensatePatternOffsets" is enabled (default), it is not necessary to set precise values for HardwareSetup - Left/Right-CalibrationPattern - Offset_X/Y/Z
- You can use SLS-2 preset or "Structured Light Setup" preset - no difference in calibration.
- To be sure that the calibration settings are correct, calibrate. Look at the green crosses, they should be placed inside the markers. Zoom in and look the small red error vectors. They indicate the difference between measure marker position and model marker position (length*10). They should be short and irregular. If they show a systematic error, there is a systematic error.
- I don't know your projector model. It may need to warm up before calibration.
- The size (diameter) of the markers is irrelevant. Only the center position is important. The size should be roughly like in the original pattern.
- If one of your markers is located wrong, that's a problem. If the entire left/right pattern is dislocated (offset), DAVID will compensate. If it is rotated, that's a problem too.

Sven
User avatar
Sven
DAVID Developer
 
Posts: 1541
Joined: Mon Jan 22, 2007 8:59 am
Location: Braunschweig, Germany

Re: Calibration Panel - left and right inner point

Postby MagWeb » Thu Jul 10, 2014 9:54 am

@ Narmella
This is a very old idea. We use it since DAVID2.x times. http://forum.david-3d.com/viewtopic.php?f=1&t=398&p=1808&hilit=Check+corner#p1808
There are so many useful hints in this forum. But how to keep them visible?
I'd like to ask anybody for help indexing important posts here:
http://forum.david-3d.com/viewtopic.php?f=7&t=1118

@Tibor
Your gab is due to an offset of the patterns. DAVID's CompensatePatternOffsets compensates this and "moves" the virtual planes to the expected position. No need to worry about.
(Not good - calibration might fail - but possible: Disable CompensatePatternOffsets to keep the virtual edges fitting together )
Think the result posted above by sts are pretty near to be ideal.
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Calibration Panel - left and right inner point

Postby sts » Thu Jul 10, 2014 11:33 am

Sven wrote:Hi STS,

Gunter has already given great tipps and answered your questions, correctly of course, I just want to confirm officially:

- You don't need to change any Advanced Settings unless you build a completely different calibration corner.
- It does not matter if the red and green distances (your first post) have the same length. They don't. DAVID can compensate such offsets ("CompensatePatternOffsets"). It's just not possible to produce foldable perfect calibration corners at a reasonable price. Also when you use the back side (240mm) the offsets will be different anyway. That's why we made DAVID compensate that.
- When you build your own calibration corner, make sure the patterns have no error in ROTATION.
- "ExpectedPlaneOffsetsOfOuterCorner" is for "negative corners" only, not relevant here.
- When "CompensatePatternOffsets" is enabled (default), it is not necessary to set precise values for HardwareSetup - Left/Right-CalibrationPattern - Offset_X/Y/Z
- You can use SLS-2 preset or "Structured Light Setup" preset - no difference in calibration.
- To be sure that the calibration settings are correct, calibrate. Look at the green crosses, they should be placed inside the markers. Zoom in and look the small red error vectors. They indicate the difference between measure marker position and model marker position (length*10). They should be short and irregular. If they show a systematic error, there is a systematic error.
- I don't know your projector model. It may need to warm up before calibration.
- The size (diameter) of the markers is irrelevant. Only the center position is important. The size should be roughly like in the original pattern.
- If one of your markers is located wrong, that's a problem. If the entire left/right pattern is dislocated (offset), DAVID will compensate. If it is rotated, that's a problem too.

Sven


Thank you Gunter, thank you Sven!

- I've now resetted the advanced settings, because i don't need to make changes for my calib panel.
- I use the Acer projecter that comes with det SLS2-Set. And i let the projector at least 15 min running before using it for scanning and/or calibration.
- The rotation of the pattern and the 90° angle of the two panels to each other sould be no problem.
- As soon i can get time on the cnc mill, i'll make a panel with a corrected pattern.

If the panel is ready i think, that my problem is gone.

I'll do both checks on the calibration:
- Sven: Check all green crosses and looking for short and irregular small red error vectors.
- MagWeb: Scanning the panel after calibration and display both Calibration Panel and Scan in SL-Mode.

sts
sts
 
Posts: 16
Joined: Fri Jun 27, 2014 1:35 pm

Re: Calibration Panel - left and right inner point

Postby sts » Thu Jul 10, 2014 2:58 pm

Sooooo... a cnc session, a calibrating session, a scan session and a (short) alignment session. And it looks decent:
pipe_done.png

(I have not placed much value on the inner part of the pipe)

The Calibration:
2014-07-10 14_39_05-DAVID-LASERSCANNER v3.10.0.jpg

x2014-07-10 14_39_31-DAVID-LASERSCANNER v3.10.0.jpg


The both right is the largest red lines. But totally other direction.

sts
sts
 
Posts: 16
Joined: Fri Jun 27, 2014 1:35 pm

Re: Calibration Panel - left and right inner point

Postby sts » Thu Jul 10, 2014 3:05 pm

The "MagWeb"-Check:
y2014-07-10 14_41_34-DAVID-LASERSCANNER v3.10.0.jpg


Overall it looks nice.

Again the result but with texture:
pipe_done.png


Thank you to all.

sts
sts
 
Posts: 16
Joined: Fri Jun 27, 2014 1:35 pm

Re: Calibration Panel - left and right inner point

Postby BoneIndustries » Thu Jul 31, 2014 6:15 pm

sts wrote:Again the result but with texture:
pipe_done.png

sts


What did you use for the red dots? Stickers or paint? How many dots did you pick up in each scan? I would imagine you would want to have at least three dots to make a triangle in a plane of locating. I have a wide thin part that is having a hard time aligning scans when I go from one side to the other. I'll try to post images later.
BoneIndustries
 
Posts: 43
Joined: Wed Nov 20, 2013 6:18 am

Re: Calibration Panel - left and right inner point

Postby BoneIndustries » Fri Aug 01, 2014 2:49 pm

Can someone please let me know what geometry of the calibration panels that cam with the SLS-1 kit looks like. The plastic hinges have come unstuck and I think that I am seeing bowing of scans.

Also, I measured the distance from the corner edge to the first dot on each panel and there is a difference of 2mm. I'll post a pic later when I get home.
BoneIndustries
 
Posts: 43
Joined: Wed Nov 20, 2013 6:18 am

Re: Calibration Panel - left and right inner point

Postby BoneIndustries » Fri Aug 01, 2014 9:30 pm

Below is how my calibration corner came with my SLS-1 kit. Even thou two of the hinges became unstuck, I was able to re-stick them in the same location, allowing for the two pieces to just just touch in the corner. (The top corner bracket is removed for clarity.)
20140801_160239.jpg


In the two following images you can see that the distance from the corner to the reference mark for the overall location of the dots are different. The left side measures 3/16" (4.76mm) and the right side measures 3/32" (2.38mm), for a difference of 3/32"(2.38mm).

Left side from the top
20140801_160519.jpg

Right side from the top
20140801_160604.jpg


What needs to be done to get the calibration corner correct? I always thought something was a little off with my scans, but I couldn't figure out what is was until I started reading this thread.
BoneIndustries
 
Posts: 43
Joined: Wed Nov 20, 2013 6:18 am

Re: Calibration Panel - left and right inner point

Postby BoneIndustries » Fri Aug 01, 2014 10:28 pm

Here is a part and scan that I working on.

Exposure: 1/60
Brightness: 159
Quality Check: 0.4
Auto Texturing enabled
Auto align after each scan, then preform Global alignment.
20140801_171457.jpg
2014-08-01_17-15-19.png
BoneIndustries
 
Posts: 43
Joined: Wed Nov 20, 2013 6:18 am

PreviousNext

Return to Camera Calibration

Who is online

Users browsing this forum: No registered users and 1 guest

cron