RTIs of dark stone

[Anna Serotta]

Hello all,

I am currently working on a project that involves documenting tool marks on Egyptian stone sculpture with RTI.  Although I'm generally getting very useful images this way, I'm having a particularly hard time with dark stones like diorite, basalt, granodiorite, etc.  With these stones I often get strange gray/black patches on areas of the image that almost look like translucent overlays (see attached screenshot of RTI Viewer with image of granodiorite sculture).  These patches sometimes appear in areas that are in shadow, but often not.  Also, I have not see this phenomenon on RTIs of sculpture made from lighter colored stone.  I suspect it might have something to do with the mix of dark, reflective minerals and semi-translucent minerals, but I'm not sure.  Has anyone experienced similar issues, and if so, did you figure out a work around?  The image looks significantly better in other rendering modes, but I want to make sure that the RTI I'm generating contains accurate data, and that the normals aren't somehow being compromised.  

Thanks for your help!

Best,

Anna 

Anna Serotta

contract conservator

The Sherman Fairchild Center for Objects Conservation

The Metropolitan Museum of Art

anna.s...@metmuseum.org

[João Barbosa]

Hi Anna,

   What format are you using? PTM of HSH?

Joao Barbosa  
University of Minho 
Computer Science Department

[Lindsay MacDonald]

Hi Anna, I have had problems in the past using the old PTM fitter
on areas that are dark or in shadow. The problem seems to be that
the algorithm produces coefficients for the biquadratic function
such that it has no maximum. (If this is the problem you can see it
gives the appearance of 'posterised colour' in the normals image.)
This confuses the reconstruction algorithm in the Viewer.
I was able to improve the performance in my Matlab software
implementation by taking a subset of the histogram of intensity
values at each pixel, thereby excluding the illumination angles
where it was in shadow. But you can't do that in the standard Fitter.

I would suggest that you pre-process your image set using
batch mode in Photoshop to lighten the shadows, or somehow
modify the image tone curve, before starting the Fitter.

Lindsay MacDonald
Dept of Geomatic Engineering
University College London

--
Group website is: https://sites.google.com/site/buildrtis/
---
You received this message because you are subscribed to the Google Groups "RTI_help" group.
To post to this group, send email to rti_...@googlegroups.com.
To unsubscribe from this group, send email to rti_help+u...@googlegroups.com.
 
 

[Anna Serotta]

Joao, I was using PTM fitter originally.  I tried re-processing the data set with HSH fitter and also lightening the shadows slightly in Photoshop as Lindsay suggested, and combination of both of those changes gave me a much better, more useful RTI.  

Thank you both!