Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - FLuca

Pages: 1 2 [3] 4
Face and Body Scanning / Re: wire or wireless Remote Switch?
« on: December 15, 2014, 12:59:03 PM »
hi polygon
i have developed both solutions to trigger 16 cams and the wire if more reliable so far.
my wireless system was based on radio triggers from adafruit, and depending on the remote control location i have delays and black frames
maybe with expensive triggering systems from the market, you can sync with more reliability
feedbacks would be appreciated

Feature Requests / Re: Slicer
« on: September 29, 2014, 04:30:05 PM »
"Hide selected points" would be nice to help cleaning point cloud before mesh generation.

Face and Body Scanning / Re: Most predicable set for better alignments
« on: September 10, 2014, 03:47:25 PM »
Too bad !
So what is the purpose of gradual selection for the standard version ?

Face and Body Scanning / Re: Most predicable set for better alignments
« on: September 10, 2014, 11:59:18 AM »
Hi Digitage,
I use PS for many months now with acceptable results but I think I have missed something.
How do you optimise the cam alignment ?
Is it a feature of the pro version?
In the standard version I have not found such feature.
Do you mean selecting cameras and re-applying 'Align Selected cameras' ?
I do this systematically and apparently it does move the cam slightly so I suggest they are optimised. But with this procedure, all points removed manually from the sparse cloud seems to be used in the alignement as they appear again in the new spare cloud. Therefore I do not see the point in removing points with gradual selection.
Could you please detail a bit more your optimisation step ?

General / Re: How much does colour difference have an impact on PhotoScan ?
« on: September 09, 2014, 03:42:45 PM »
Hi Exhale,
normally you should have consistent colour temperature with your cam in manual mode, or at least with this particular setting being fixed. shooting in automatic mode will give you inconsistent colours and lighting appearance, which will decrease matching process efficiency.
to get matching colours over a set of images, you can use raw images from you cam and set the colour temperature with lightroom or any other software processing the raws.
you can also change the colour temperature from any other picture type (jpg, png, tiff) with lightroom but I guess you will have to adjust it manually for each image
using raw images allows a wide range of adjustements after the shooting which are not possible with jpg (or with less control)

Face and Body Scanning / Re: Triggering and Camera Questions
« on: September 02, 2014, 03:52:39 PM »
Regarding the noise issue, i have allready bought raw models from infinite, and they are quite noisy even with 110 cams. Infinite advantage with such quantity of cams is to have a great coverage of the subject. He even show himself how to get rid of noise with zbrush in its famous tutorials.

From my modest experience, noise quantity really depends on the subject and its skin quality.
With children with very smooth and spotless skins, the result is very noisy with big holes in the point cloud, while with old skin full of moles and wrikles the result are much more precise.

You can also notice the quality difference between skin and clothes. The later are always better defined due to their more detailed textures (with mat fabric of course).

It is not a mystery that PS is requiring images with lots of distinctive features and skin is not the ideal material (and can be very shinny). Hair are even worst.

It is also very important to have the largest depth of field possible (high aperture number thus lots of light) as blur on the subject is very detrimental for the model quality.

Good luck with your rig

Feature Requests / Re: PSD Texture Mask Collage
« on: August 28, 2014, 04:44:01 PM »

this option would be very usefull, it would allow us to blend manually the different images like it is possible in photo stitching programs such as realviz or autopanogiga

Hi Kms1,
I have only 15 DSLR but I have good results for facial scan with the attached config.
With 15 I can only get the front face in one shoot, and hair are very messy and need complete rescult.
I think with 20 you can get full 360 as you don't need so much cams on the back which has less occlusions.
Neutral powder help reducing the specular points on skin and even on hair.
First scans are time consuming, especially the USB setup that can be tricky to be stable.
Even if individually all cams can be connected to the computer, you start to have stability problems with all cams together. I would recommend usb3 cables, even if cams and hub are usb2.

I wish you good scan and show back your results !

Face and Body Scanning / Re: Turntable based Full body capture rig
« on: June 30, 2014, 05:33:03 PM »
Very nice work Mark !

How do you deal with the "droit ? l'image" issue with your models that are displayed and sold ?

Face and Body Scanning / Re: Scanning birds or stuffed animals
« on: June 23, 2014, 10:48:04 PM »
did anyone tried polarizing filters?
I understood from previous posts that it could help with hair but never tried it yet
problem is cost with multiple cam rig but if it save time it may worth it

Face and Body Scanning / Re: moving mesh inside Agisoft PhotoScan Pro
« on: June 04, 2014, 12:28:33 PM »
you have to edit the original xml file with a text editor and replace each line of the camera transforms matrices by the new ones included in the new_cams.txt
be careful to replace the correct camera transform line

Face and Body Scanning / Re: moving mesh inside Agisoft PhotoScan Pro
« on: June 03, 2014, 03:59:37 PM »
export you low resolution model in FBX format to have cameras in your scene
after moving and scaling, you can then export the camera transform matrices with the attached script

Face and Body Scanning / Re: moving mesh inside Agisoft PhotoScan Pro
« on: June 03, 2014, 12:00:15 PM »
Hi Claude
it is possible to move and even scale a low version of the model with the cameras in an external application, you can then extract the transform matrices of the cameras and update them in the original exported cam.xml file
when you re-import the modified cam.xml, you can re-generate a high version of your model, with the desired scale and position
I do it with 3dsmax and it works perfectly

It is a good idea to share configurations, as it seems that everybody it developping its own control system.
I have now a rig with 15 canon 600D that works quite fine and I hope will grow with time.

I use smartshooter for camera settings and automatic download of images to the computer.
It works very well but the hard part is to get all cams with stable usb connection to the computer.
As already said in the forum, good cables are required. usb 3 cables have a better screen and allow more length than usb 2 cables.

for the cam triggering I don't use smartshooter because of the delay between cams which increase with each additional cam. it works if you are in an almost black room with the shutter open a very long time (2-5s for 10cams) and trigger manually the flash when you have eard that all mirrors ar up.
possible but not very convenient and very un-confortable for the model that go from full black to big flash lights

instead I have developped a small electronic board piloted by an arduino that drive relays. contacts of relays are wired to the remote control of each camera.
one additional relay is driving the flash remote control. that way it is easy to automatically fire the cam and the flash after a predefined timer (120ms in my case).
connected to the arduino, a wireless remote control from adafruit allow firing from where you want.

with this config I manage to get sync at 1/5s without any black images

the bottleneck is the download of pictures to the computer that takes between 20s and 30s (for 15 pics) even with a SSD. it is a very long time when searching for a nice pose

Face and Body Scanning / Re: Mounting cameras to light stands
« on: May 13, 2014, 02:30:21 PM »
with the manfrotto super clamp you need to add a spigot  adapter and a basic rotule head
for spigot it is preferable to have a long one ( to allow your camera to be more freely oriented
for the rotule you can use the Manfrotto MA492 which doesnt have a removable plate but have a very small footprint allowing to open the battery case without removing the cam
to mouunt the clamp on you can also use the Manfrotto autopoles which can avoid the large space taken by the tripods, three heights are available. they can also be mounted on a tripod if your ceiling is too high


Pages: 1 2 [3] 4