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

Scanning of calibration panel

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

Re: Scanning of calibration panel

Postby argo » Tue Nov 12, 2013 3:42 pm

Lunat1c wrote:-If you also want to get rid of the little pixel like reliefs (they're literally pixels of your projector), you can put the board a little out of focus of the projector. This way you have a smoother scan, but you lose sharpness. But I don't prefer it, I'm fine with those little imperfections as far as the scans are sharp.


Thank you for the tip. but i think to go for the sharpness same as you.
gkrywusha wrote:with the ability to really control brightness, sharpness, contrast and gain my black/white grid is strong. And never going autofocus was a must. Also, using a 4+ lens on my projector allowed a much better clarity of the grid.

As much as we have the same camera, we wont be able to get the same results if we owe different projectors.
In my case the problem is about the synchronization of webcam and projector! and as far as i tested, there is only one thing i can do to solve!! this issue. and that is placing the webcam right above the projector. and the last scan on this topic was made by this way.
we really should be careful to choose the right projectors.
and i am going to buy another one but i have no idea which brand, which model, i need to buy!!

Lunat1c wrote:when I forgot to deactivate autofocus, I got this

At the first look, it is quiet similar to my first post here. but i have a modified webcam. which means the lens and all the associated are removed. and there is no way to have the autofocus ability.

Thank you both for the reply. and good luck
Shahram,
User avatar
argo
 
Posts: 141
Joined: Sun Jul 14, 2013 8:48 am

Re: Scanning of calibration panel

Postby argo » Sun Dec 01, 2013 6:23 pm

Hello,

What is the reason of these dots every where? light? camera quality? or something else?
is this calibration ok?
Please mention it, if you see any thing on this image that can affect the scans.

Regards,
Attachments
2013-12-01_20-36-07.png
default patterns
2013-12-01_20-36-18.png
User avatar
argo
 
Posts: 141
Joined: Sun Jul 14, 2013 8:48 am

Re: Scanning of calibration panel

Postby MagWeb » Mon Dec 02, 2013 12:57 am

Maybe your image is just a bit bright to get these dots (but don´t worry your result is quite good).
Another thing (DAVid allows this but you might get better results avoiding it): On the righthand you captured some area out of the corner.
Using a "normal" inner corner forces reflections to the opposite plane. You may try a "negative corner"...
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Scanning of calibration panel

Postby argo » Mon Dec 02, 2013 3:44 pm

Hi Gunter,

Thank you for the tip (warning) about those empty area.
i got a negative corner, and the result not much different from before. I guess it is all about the cam and projector synchronization. and have no idea how to solve that.

2013-12-02_17-52-05.png


anyway i was wondering why i can not scan this little stone statue. as far as i checked it is not a shiny object and it is totally white. and that was driving me mad... today i applied this white washable spray and got an amazing result to compare with previous ones. so this david scanner is like a college and we need to learn and study more about the nature of objects. :roll:
(what is the reason of those horizontal lines on the right image? synchronization problem?)

Untitled-1.png
left(before whitening) right(After)


Regards
User avatar
argo
 
Posts: 141
Joined: Sun Jul 14, 2013 8:48 am

Re: Scanning of calibration panel

Postby MagWeb » Mon Dec 02, 2013 8:28 pm

As before, your cornerscan is definitely o.k. You may try to get an even better result increasing the number of phaseshifts (AdvancedSetting/StructuredLight/Paseshifts/Shifts).

You said your object is of stone - maybe the stone owns some transparent particels?

at the small waves:
These seems to appeare V-shaped? They seem to start at the chin and expand to the cheaks while the top lip is ok. This V effect should be caused by the rolling shutter of your Logitech CMOS sensor. This gives a uppwards moving darker zone in the cam image. The result on a 3D object are such V-shaped wave zones.
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Scanning of calibration panel

Postby argo » Tue Dec 03, 2013 8:35 am

Hi Gunter,

Thanks you so much for all your help.
As much as i hate but i think it is the time to use the "give up" words.
I did try every possible setup/setting (all refresh rates with all exposure time, everything related in advanced setting, different position of camera, different lens, low and high light of projector, positive and negative corners, .......) but i could not get ride of these horizontal lines( you say V-shape?)

2013-12-03_10-46-43.png


I know it has not anything to do with david and the problem is because of my hardware(s). but which one? cam? projector? graphic card? chair? my hair style? :roll:

Regards,
User avatar
argo
 
Posts: 141
Joined: Sun Jul 14, 2013 8:48 am

Re: Scanning of calibration panel

Postby hunkatibor2 » Tue Dec 03, 2013 8:51 am

Hi
I also have a similar mark in my, did I just I use a laser.
Tibor
Attachments
2013-12-02_17-48-11.png
Paneles
2013-11-28_18-41-56.png
2013-11-28_18-08-56.png
48 mm
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: Scanning of calibration panel

Postby MagWeb » Tue Dec 03, 2013 11:48 am

Hi Tibor and Argo,

the specific problems of SL - and Laserscaning are different things, but DAVID uses the same source in both cases: The cam.
In both cases DAVID compares image frames assuming never changing ideal conditions where the cam captures a scene that does not change at all.
In freehand laserscanning rapid motion of the laser may give some smearing of the line's image, the laser's line may not be straight enough and so on...

Back to SL (Argo forget that V-shaped thing your latest image clearly shows something different):
My latest attempts show that these horizontal waves are due to the interference of two devices (cam and projector) running asynchronously.
Asynchrony results in a cam image stream that changes its brightness. This makes DAVID calculate waves also.
Only a few cams give the option to influence the frequency (pixel clock) of the cam (e.g the DAVID-3-M does, also uEye...).

To get an idea about your system's synchrony yo might try:
Set the second monitor's background (projector) to a medium grey;
Project onto a flat white surface.
Start DAVID. Set the projection to monitor 1 (or minimize it);
Go to DAVIDs calibration mode and watch the red line (which should be a curve in case of scanning). It should be a pretty horizontal line now, for the cam only watches the plane grey projection. It may be some noicy line (this doesn't matter here) but it is most important that this line in total does not move upwards/downwards during a time which is needed to do a scan (zoom in to see this better).
Try possible framerates and exposure settings if there is some movement. Is there an exposure value that stops this overall motion?
If yes: This is the exposure to be used (you should prefer changimg the images' contrast and brightness via iris or other driver settings).
If no: Change the frequency of your projector (note: changing its resolution or connection type(HDMI, DVI, VGA) may give you different possibilities).
Now again:
Try different framerates and exposure settings if there is some movement. Is there an exposure value that stops this overall motion?
.... again and again. Never give up till you walked through all possible combinations.

Doing this I found a set of settings running my 720p projector (@60 or 50Hz) at a smaller resolution of 800x600 (@56Hz) that gives a awesome smooth result.

EDIT:
Argo: Watching your latest corner results over all: These are already pretty good - for my feeling the BW contrast (= the waves amplitude) could be higher (this will make sure to get the projection even on darker surfaces). Our hardware, I think also the most expensive one, will always have some errors (at least simply to the fact that WINDOWS and also USB do not work in real time). We can only try to find ways to minish them to a quality that fits to our needs.
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Scanning of calibration panel

Postby argo » Tue Dec 03, 2013 4:24 pm

Hi Gunter,

Thank you so much for the time you put to explain and following the problem.
no matter if my system gets to work or not, i do appreciate your help.

I went through your instruction and i did all the tests once via VGA and once via DVI.
there was no different between VGA and DVI at all.
i got the 60 hertz is the best value (the difference of 60 to compare with (56 or 70) was not too much)
I got the best exposure time is 1/30 (this one also not much different with 1/15 but much better than 1/60)
The thing is i could not see a straight horizontal line all the time with any setting.
I zoomed at a part and made it fix then there was other part of line flicking.

Untitled-1.png


I think you are right about the sensor of my cam. it is like a circle and center of image is different from corners. and as much as we get far from the center we see more darkness.
Are all the webcams like this?
there was only 2 options to make the lines fixed and straight. :1- using the zoom option(driver zoom) 2- reduce the webcam contrast (nearly zero)
just mentioned these ways to show you how far i went to check the setting.

Can you please upload your panel scan(zoomed and not zoomed)? i would like to see the differences.

Regards,
User avatar
argo
 
Posts: 141
Joined: Sun Jul 14, 2013 8:48 am

Re: Scanning of calibration panel

Postby MagWeb » Wed Dec 04, 2013 12:44 am

Hi Argo,

thanks for checking.
IMO your brightness falloff towards the frames edges isn´t a a real problem. It may be caused by the projector (spreading the light in a not even way) or by the optics fixed on a your Logitech (if that`s the reason maybe you could solve this vignete effect spending some money on optics made for a larger sensor). Anyhow a falloff will be always there projecting on 3D surfaces as the calibration corner is (due to the changing distances and reflection angles). DAVID tries to compensate this calibrating the system...
Seems the either point is whether such a line moves (means different cam-frames show different brightnesses of the same projector-lightened scene). Depending on your hardware and its driver's possibilities there may be some limits.

Following the problem isn`t some problem for me for it isn't "your" problem. Its simply the question "how to get the best input that DAVID-SL is able to decode". We need a workflow to judge the limits and possibilities of our different setups. Maybe there are some other ideas than mine out there?

Here are my latest corner results @ default settings + shifts set to 16:
Attachments
2013-12-03_22-33-12.png
2013-12-03_22-33-45.png
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Scanning of calibration panel

Postby argo » Wed Dec 04, 2013 6:59 pm

Gunter.

This becomes kind of my dream to have smooth scan like yours.
I can have good scans at the final but they take too much times for cleaning. that is why i am looking for the improvement. and i dont want to waste the time simply just for cleaning.
Usually how much time do you spend generally to clean e.g 8 scans? for me it takes average of 2hours to check all overlap parts!
here is my new panel and i used separate prints that is why there is a space at the center of panel. is that ok? or i need to change it?

Regards,
Attachments
2013-12-04_21-11-29.png
scan with normal setting + shifts set to 16
2013-12-04_21-27-00.png
User avatar
argo
 
Posts: 141
Joined: Sun Jul 14, 2013 8:48 am

Re: Scanning of calibration panel

Postby MagWeb » Thu Dec 05, 2013 1:11 am

:) Thanks - Maybe you allready visited my ugly user page at the DAVID-Wiki? Over there you can watch my very first partial scan in early 2007. The dream of improving this kept me going all the time.

@ time to clean a set of scans:
Hmm. hard to say... depends very much on the surface and the quatity you've in mind...
With a reliable calibration:
To keep cleaning work controlled I start with a lowest-res-fusion (or a guessed minimal resolution) after aligning without any cleaning.
Watching the result:
Are there some disturbing overlaps? If yes, - in most cases no-, clip them switching back to the partial scans.
Do I nead more details? If yes do the next higher fusion
Watching the next result:
Are there some disturbing overlaps? If yes, - now there may be some-, clip them.
Do I nead more details? If yes do the next higher fusion.
a.s.o.
This way it may happen that I reached the wished level of detail without any cleaning. Outliers will be isolated blobs and these are easily removed via SelectConnected at a fusion reult.
Nevertheless: If I decide to do some bigger cleaning I export the aligned partial scans and import them to MeshMixer. This freeware gives a bundle of options for cleaning (this should be discussed in another thread...we' re in the Calibration section). After this work ofer there: reimpot and fuse...
MagWeb
Moderator
 
Posts: 2499
Joined: Wed Jul 18, 2007 8:48 pm

Re: Scanning of calibration panel

Postby hunkatibor2 » Thu Dec 05, 2013 9:35 am

Hi Gunter
You have the MeshMixer, unfortunately I always hang of what version of what you use?
I propose a graphical result of the calibration software and evaluate the results can be saved, sent to members of the forum, so the results would be comparable.
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: Scanning of calibration panel

Postby argo » Thu Dec 05, 2013 10:11 am

Thank you Gunter for the tips for fusion part.

hunkatibor2 wrote:I propose a graphical result of the calibration software and evaluate the results can be saved, sent to members of the forum, so the results would be comparable.


Tibor,
I think a video clip of these steps is really helpful for members: 1-placing the object in front of the scanner, 2- focusing both camera and projector 3- replacing the panel and calibrate and scan it.
but we need to see both views of david and real object at the same time.
Hope Sven makes this kind of tutorial soon.

Regards,
User avatar
argo
 
Posts: 141
Joined: Sun Jul 14, 2013 8:48 am

Re: Scanning of calibration panel

Postby argo » Thu Dec 05, 2013 11:49 am

Gunter,

In your picture i think your camera is little bit out of focus. am i right?
Because i can have a little smoother scan when i focus away my camera.

2013-12-05_14-08-44.png


here is what i do. place the object then well focus both camera and projector. remove the object and place the calibration panel. so part of calibration panel (should be middle of the panel?) will be at the focused area. am i doing fine?

Right now i am like those students whom get more confusion by more study :?

Best Regards,
User avatar
argo
 
Posts: 141
Joined: Sun Jul 14, 2013 8:48 am

PreviousNext

Return to Camera Calibration

Who is online

Users browsing this forum: No registered users and 1 guest

cron