Agisoft Metashape

Agisoft Metashape => Face and Body Scanning => Topic started by: Xilstudio on January 27, 2013, 07:44:55 PM

Title: Would this work? Video, turntable solution
Post by: Xilstudio on January 27, 2013, 07:44:55 PM
Would this work?
I have a wooden lazy susan, rigged up with a simple pulley.  You pull the rope and it unwinds as you pull.  works, cheap, easy.
What if placed a human model on this, set up for say 4 or 6 video cameras set for HD.  you record the rotation... you export the films as jpg sequences, choose the frames you want.
I know there would be quality loss since I would not be using 70 DLSRs, but would it function?  I am trying to a balance between speed, cost and quality.
Title: Re: Would this work? Video, turntable solution
Post by: Infinite on January 28, 2013, 03:13:15 PM
Would this work?
I have a wooden lazy susan, rigged up with a simple pulley.  You pull the rope and it unwinds as you pull.  works, cheap, easy.
What if placed a human model on this, set up for say 4 or 6 video cameras set for HD.  you record the rotation... you export the films as jpg sequences, choose the frames you want.
I know there would be quality loss since I would not be using 70 DLSRs, but would it function?  I am trying to a balance between speed, cost and quality.

You really need to record uncompressed video. The video compression on most consumer cameras is very poor. Even at HD, also 24fps is undesirable. The only benefit are the lens ranges available for DSLR's
Title: Re: Would this work? Video, turntable solution
Post by: RalfH on January 28, 2013, 03:22:31 PM
As has been said before in this forum, video is a poor source of data for structure-from-motion approaches unless you are working with a high-end camera which (1) has very high true shutter speeds and (2) allows you to save uncompressed video.

Also, in the described setup it would be extremely difficult to find synchronous frames in the 4 to 6 video sequences. If there is any movement of your model (except for the rotation), everything will be messed up. Living people breathe, for example.
Title: Re: Would this work? Video, turntable solution
Post by: andy_s on February 01, 2013, 06:56:15 PM
Does anyone know of / comment on available [to mere mortals] "non-rigid registration" type software that may help ? eg create 3 high quality photoscan pointclouds [distributed in time (and therefore some deformation)] then "stitch" them together with said software ?

See - http://www.youtube.com/user/haoli81 - only if you can spare 64minutes  :)
Title: Re: Would this work? Video, turntable solution
Post by: RalfH on February 01, 2013, 07:05:57 PM
You mean this stuff? (http://www.mpi-inf.mpg.de/resources/deformableShapeMatching/EG2011_Tutorial/slides/2.3%20Non-Rigid%20Registration.pdf (http://www.mpi-inf.mpg.de/resources/deformableShapeMatching/EG2011_Tutorial/slides/2.3%20Non-Rigid%20Registration.pdf)) Ask the author of that presentation.

But why would you want to do this? If all you want to do is create a 3D model of a living person, using multiple synchronous cameras is much easier, at least if that person comes into your studio and agrees to have such pictures taken.
Title: Re: Would this work? Video, turntable solution
Post by: andy_s on February 01, 2013, 08:16:09 PM
Thx RalfH - that's a good slide summary of the video. It's an economic comparison i'm considering: less cameras v more time/deformation xtra software... 
Title: Re: Would this work? Video, turntable solution
Post by: RalfH on February 01, 2013, 08:38:21 PM
andy_s,

"time is money" could be a counter-argument. My opinion is that you'd be much better off using a number of low-cost (even used) point-and-shoot cameras with which you can shoot synchronously (see, for example: http://stereo.jpn.org/eng/sdm/index.htm (http://stereo.jpn.org/eng/sdm/index.htm)). And  20 used point-and-shoots are probably cheaper and have higher resolution and better image quality than 6 good video cameras.
Title: Re: Would this work? Video, turntable solution
Post by: andy_s on February 01, 2013, 09:54:59 PM
Appreciate that Ralf.

My lower limit is 18MPix

The compare I have in mind is [say] 60 cam instant capture v [say] -

16 cam [arranged in 4 groups of 4 at 22.5 degrees to each other] then

For i=1 to 4
capture point-cloud, rotate turntable 90 degrees
next i.

[say] 10 second total capture.

Stitch 4 [overlapping] point clouds together with said non-rigid registration software.

Not sure the concept is sound but i'd like to investigate further. 
Title: Re: Would this work? Video, turntable solution
Post by: RalfH on February 01, 2013, 11:08:33 PM
andy_s,

don't know what you're aiming at - in the beginning youtalked about HD video cameras (which have approx. 2 MPix), now you say your minimum is 18 MPix?

If whatever it is that sits on your turntable is alive (e.g., a person), you'll have problems. Mislaignments and/or ugly model erros. With anything that lives, synchonisation is the way to do it. If it is just objects, that setup might work. But make sure that your rotation angles are not multiples of your angular camera spacings or you'll have identical positions with different cameras (i.e., no additional information).

Title: Re: Would this work? Video, turntable solution
Post by: andy_s on February 01, 2013, 11:39:14 PM
Hi Ralf,

it was actually Xilstudio who referred to HD Video.

It is the "problems" that you refer to that I would like to investigate whether the "non-rigid registration" software would greatly [and hopefully economically] reduce.

I appreciate your further comments.
Title: Re: Would this work? Video, turntable solution
Post by: RalfH on February 01, 2013, 11:41:41 PM
andy_s,

sorry, had overlooked that...
Title: Re: Would this work? Video, turntable solution
Post by: andy_s on February 02, 2013, 12:01:05 AM
No need to apologise Ralf - very easy mistake to make.
Title: Re: Would this work? Video, turntable solution
Post by: RalfH on February 02, 2013, 01:21:27 AM
I only know the "non-rigid registration" stuff from a quick search after your post, but to me it sounds complicated and work-intensive (and potentially error-prone), so personally I'd stick to a proven method for now.
Title: Re: Would this work? Video, turntable solution
Post by: andy_s on February 02, 2013, 02:42:37 AM
This maybe useful:

http://www.pcl-users.org/Different-geometeies-and-non-rigid-3D-Point-Cloud-Registration-td3766532.html

I wonder if they did get it into Meshlab yet:

http://www.pcl-users.org/Adding-Point-Cloud-Library-to-Meshlab-in-Qt-td4022240.html
Title: Re: Would this work? Video, turntable solution
Post by: hsmith on February 02, 2013, 03:30:17 AM
Andy,

I think your suggestion about using non-rigid registration to align meshes that you get from scans using less than 10 cameras is quite brilliant, as a low-cost alternative.
 
I've been corresponding with Hao Li about using his BeNTO3d software (check out http://www.bento3d.com/BeNTO3D/BeNTO3D_-_teaser.html )
He's seems pretty willing to help, but I haven't given him very good sets of data yet.
Still working on my rig.

Have you tried some of the non-rigid registration programs that have been developed for medical research?  3D Slicer seems the most user-friendly, (http://www.slicer.org/ )
but there are also a few others, like elastix
( http://elastix.isi.uu.nl/index.php. )

Keep up the great experimentation!

Harold
Title: Re: Would this work? Video, turntable solution
Post by: Infinite on February 02, 2013, 04:06:19 AM
This might help - http://www.ir-ltd.net/scan-of-a-scan (research I was doing in 2010)

FYI Artec Group software has a basic version of Hao Li's Non Rigid Registration code integrated into it - http://www.artec3d.com/shop/

It took me a couple of years (meetings with Hao Li) and many emails but I got the two parties to work with each other, thus improving Artec scan quality whilst giving other users the ability to utilize the code.
Title: Re: Would this work? Video, turntable solution
Post by: andy_s on February 02, 2013, 01:06:11 PM
Harold, Lee,

wow - the grey sky just turned blue, THANKYOU guys !  :) :)

@Lee - I have a vague memory of looking at that on your site a few weeks ago [when I first discovered "photogrammetry"] - it musn't have "registered". I figured Artec must have implemented Hao'Li's [or similar] algorithm - it's what started the recent discussion on this thread. You are a true inspiration !

@Harold - i'll check your links out later but just need to go and do a 360 pirouette in the kitchen [in 1 second] first  ;). I just checked and have got slicer "bookmarked" but have promptly forgotten about it.
Current euphoria is blinding, getting ahead of myself now - I wonder if a 4 vertical camera system / 16 photoscan point-cloud / 1 second [or at least very low] capture / rotating external rig [not so dizzy!] / Hao-Li "donkey work" [don't ever let him see that description  ;)] solution is possible...

Wish I knew how cameras worked - anyone a decent link [with timings] to explain how image goes from Non Volatile Cache memory [and what size is it] -> Non Volatile RAM -> card reader OR usb and when is cache cleared [that's just me making things up as I think about it so apologies if its total twaddle]  :-[ 
Title: Re: Would this work? Video, turntable solution
Post by: hsmith on February 02, 2013, 06:42:32 PM
I know exactly how you feel, Andy, about the euphoria.  Lee, you have certainly been an inspiration for all of us, especially how you've figured out how to overcome the many technical problems associated with live scans.  Interesting about Hao Li's work with Arctec, although it's still a bit too expensive for my budget. 

Speaking of problems of of the lazy-susan, I'm planning to use a camera dolly on a circular track that goes around the model, always pointing to the center to capture a (as still as possible) model.  I'll let you know how that works.

I'm using 6 Canon SX130s (got refurbished ones for about $120 each on ebay, sometimes they're cheaper).  With a lot of overlap, I'm happy with the results of inanimate models.

Sync is a challenge; I've been fairly successful using the CHDK, which most Canon point-and-shoot cameras are able to use as a multiple camera sync hack.  http://chdk.wikia.com/wiki/CHDK
Using USB cables, I can get the 6 shutters to remain open for 1/10 second, and now I'm working on some circuits to fire the studio strobes during that 1/10 second.

BTW, Andy, there are some useful posts in their forum about writing a motion trigger for lighting capture that discusses the issues you raise.  Not sure about applicability to DSLRs, however.

It's great to know there are others working on the same challenges!  Looking forward to further collaboration.

Harold
Title: Re: Would this work? Video, turntable solution
Post by: Magnus on February 04, 2013, 01:32:30 AM
Hello all!

Harold, interesting thought about dolly. I was thinking along those lines before, using something like this http://www.kesslercrane.com/kflex-s/107.htm (http://www.kesslercrane.com/kflex-s/107.htm).
Andy and I have been throwing ideas back and forth and hopefully in the not too distant future I will be able to show my build based on our exchanges.
It can maybe best be described as a wheeled, rotating light tent with cameras attached on each vertical pole, hehe. It will be manually rotating at first but later motorized.

Best, Magnus.
Title: Re: Would this work? Video, turntable solution
Post by: hsmith on February 04, 2013, 07:28:56 AM
Wow, that part about hauling the light tent around in a circle is creative!  I'm looking forward to see your build and the agisoft results.

It's amazing how bouncing ideas back and forth with other people speeds up the process.  Best of luck!

Harold