Multi-spectral RTI

[Kathryn Piquette]

Happy New Year, all! Does anyone have any recently published or forthcoming work dealing with multi-spectral RTI to report?

Cheers - Kathryn

[Taylor]

Hi Kathryn and RTI users,

I recently completed multi-spectral RTI capture and photogrammetry of a 19th century landscape painting (36 x 50 inches).  I captured 36 overlapping RTIs of the entire surface of the painting at a resolution of 500 ppi, and RTIs of certain details (for forensic identification) at resolutions up to 2,600 ppi in two infrared reflectance wavebands (715 - 950 nm and 945 - 1,150 nm) in addition to visible (325 - 645 nm) wavelengths, using a modified consumer camera.  For planning purposes, I developed a spreadsheet to calculate working distances, camera shift distances for the required overlap, and data storage requirements for the project.  My camera support setup included a 60-inch horizontal rail supported by two tripods, and a camera rotation device with a macro focusing rail.  The system was designed for portability, since the painting is located in the Boston area and I'm on the west coast (San Francisco Bay area).  Capturing overlapping RTIs provided a resolution of approximately 0.5 gigapixels for the painting, resulting in about 40 Gb of raw images before processing the RTIs.  If anyone has questions about this project, please let me know.

Best wishes for the New Year,
Taylor Bennett

[Earl G.P.]

Hi taylor

Sounds like a great project! Which fitter did you use and what hardware to process that resolution of input data? Best wishes. Graeme Earl

--
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<mailto:rti_...@googlegroups.com>.
To unsubscribe from this group, send email to rti_help+u...@googlegroups.com<mailto:rti_help+u...@googlegroups.com>.

[Taylor]

Hi Graeme,

I'm processing the images with both the PTM and HSH fitters, using the RTIBuilder from CHI (http://culturalheritageimaging.org/What_We_Offer/Downloads/).  I'm processing the RTI images on a Mac Mini (2.6 Ghz quad i7) with the Mountain Lion OS.  It's been upgraded with 16 Gb of ram, but I don't think all that ram is needed for this project.  Once the steps of the fitter have been completed, the final processing takes about 30-50 seconds per RTI.

Because of some technical glitch unique to my configuration, I was unable to run the RTI Builder on my 2009 laptop with Windows XP, although there's no reason it shouldn't be able to do the job.  I haven't yet processed the majority of the photogrammetry, although I did receive some very generous processing assistance for a subset of images using Adamtech software. 

Taylor

To unsubscribe from this group, send email to rti_help+u...@googlegroups.com<mailto:rti_help+unsubscribe@googlegroups.com>.

[Earl G.P.]

Hi Taylor,

 

Many thanks indeed. We have done some gigapan capture of RTIs on the Portus Project and encountered issues with the PTM fitter timing out. I think these were less than 0.5Gpixels but maybe not. Sounds like trying it on a Mac and also with HSH might be the solution J

 

Many thanks,

 

Graeme

To unsubscribe from this group, send email to rti_help+u...@googlegroups.com<mailto:rti_help+u...@googlegroups.com>.

--
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.
 
 

[Taylor]

Hi Graeme,

I didn't mean to convey the impression that I could process a 40 Gb RTI with 0.5 Gpixels of resolution (in less than a minute--that would be nice!).  I haven't stitched the overlapping RTI images together, I just have the individual RTIs that equate to about 0.5 Gpixels of resolution for the entire painting surface.  One purpose of doing photogrammetry is to be able to stitch the RTIs together at some point in the future; I understand it's on CHI's wish list for the Collaborative Algorithmic Rendering Engine (CARE) tools that are being developed.  My final processed .ptm and .rti files are about 1.5 to 2.5 Gb each. 

I'd be very interested to try using a Gigapan type camera arrangement for capturing RTIs of larger areas.  I'm looking at your Portus project website--fascinating work!

Best,
Taylor

To unsubscribe from this group, send email to rti_help+u...@googlegroups.com<mailto:rti_help+unsubscribe@googlegroups.com>.

--
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+...@googlegroups.com.
 
 

[Earl G.P.]

Oh I see! I didn’t notice the process timing J We are keen to process the stitched datasets and are getting closer and closer. As you say Mark and Carla have been working towards stitched and also tiled solutions (the SpiderGL PTM viewer being an example of the latter) and we also have an interest in streaming high resolution RTI datasets at this end.

 

Cheers re: Portus – we’ve been busy! All the best,

 

Graeme

 

BTW We are organising an RTI workshop as part of the CAA2013 conference on 25 March. I am arranging for an Adobe Connect conference call if anyone on this list would be interested in getting involved. Let me know.

To unsubscribe from this group, send email to rti_help+u...@googlegroups.com<mailto:rti_help+u...@googlegroups.com>.

--
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+...@googlegroups.com.
 
 

--
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.
 
 

[Lindsay MacDonald]

Hi Taylor, that's a really impressive RTI project, both the photography
and the imaging! I also found that the old PTM fitter didn't handle
large images very well. I have my own implementation of the PTM
algorithm in Matlab, and I get around the problem of large files by
processing in tiles of about 500x500 pixels.

Lindsay MacDonald
Dept. of Geomatic Engineering
University College London

[Taylor]

Thanks, Lindsay!  It sounds like your Matlab implementation could help with processing Gpixel RTI image sequences such as Graeme Earl's group is generating (above).  For high resolution RTIs of larger surfaces, your technique should be more efficient than capturing and processing many overlapping RTIs of smaller areas, as I did in my project.  My method was based on a modified camera with 16 Mpixel resolution, hence I needed many RTIs to get the overall resolution I wanted.  With larger sensors available (60 Mpixel cameras, for example), I could see how the file sizes wold become difficult to process, but it appears you're effectively dealing with this problem.

Best,
Taylor

[Earl G.P.]

Yep really useful as ever Lindsay. Perhaps we could try some of our Portus data with your fitter?

 

All the best,

 

Graeme

--

[Lindsay MacDonald]

Yes, I'd be happy to help.

One other Matlab facility I developed earlier in 2012
(I talked about it with Hembo at CAA in Southampton)
is the ability to combine multiple PTM files together
so that they can be viewed under a single interactive
lighting control in the Viewer. This can be applied even
to image sets that were taken at different times with
different cameras and processed with different Fitters.

Lindsay

[Taylor]

This sounds like a very useful application, Lindsay.  I'd like very much to try it on the RTI data sets I've acquired, if possible.  I'm also happy to share my planning spreadsheet, which can be adapted for other camera and lens combinations, with interested RTI users.

Best,
Taylor

On Sunday, January 6, 2013 3:15:17 PM UTC-8, Lindsay MacDonald wrote:

Yes, I'd be happy to help.

One other Matlab facility I developed earlier in 2012
(I talked about it with Hembo at CAA in Southampton)
is the ability to combine multiple PTM files together
so that they can be viewed under a single interactive
lighting control in the Viewer. This can be applied even
to image sets that were taken at different times with
different cameras and processed with different Fitters.

Lindsay

Yep really useful as ever Lindsay. Perhaps we could try some of our Portus data with your fitter?

 

All the best,

 

Graeme

 

From: rti_...@googlegroups.com [mailto:...@googlegroups.com] On Behalf Of Taylor

To unsubscribe from this group, send email to rti...@googlegroups.com.
 
 

[Taylor]

I've uploaded a copy of the spreadsheet I used for planning the multi-spectral RTIs and photogrammetry of a 19th century landscape painting.  The spreadsheet is accessible to members of the RTI_help group and is located in the shared folder called RTI RESOURCES.  It can be accessed using this link:  https://docs.google.com/open?id=0B1DEF8DRa0u-eV9sUEg4ekhUZDg

The file is in MS Excel (Office 2007) format.  Since Google docs doesn't display Excel files with correct formatting, it's best to download the file to view it. 

There are 2 worksheets in the file, for metric and English units, respectively.  Each worksheet is formatted to print on one page (landscape letter-size).  Input parameters are entered in metric units in the first worksheet, and the spreadsheet automatically converts the distances from metric to English units for you (2nd worksheet).  Please let me know if you have trouble accessing the file or if you have any questions about how it works or suggestions for improvements. 

The spreadsheet is intended primarily for planning captures of larger surfaces with overlapping images, but it can also be used for planning captures of smaller areas and details.  It uses basic parameters for the camera and lens to calculate the working distances, vertical and horizontal camera shift distances for overlapping images, string length, minimum sphere diameter, and data storage requirements, for a given desired resolution and number of wavebands.  The spreadsheet can be easily modified for different cameras and lenses.  

Best,
Taylor

[Taylor]

A quick update:  I'll be presenting a poster/demonstration of my RTI project at the International Conference on Computational Photography (ICCP 2013) which is being held at Harvard University on April 19-21, 2013.  The poster/demonstration is titled "High-Resolution Multispectral Reflectance Transformation Imaging of a 19th Century British Painting."  The poster/demonstrations are presented from 8:30 to 10:30 am on Sunday, April 21.  For more information about the conference, you can visit http://www.iccp13.org/

Best,
Taylor

On Saturday, January 5, 2013 1:12:11 PM UTC-8, Kathryn Piquette wrote:

[Earl G.P.]

Sounds really exciting! When will it be published?

All the best. Graeme

--
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 unsubscribe from this group and stop receiving emails from it, send an email to rti_help+u...@googlegroups.com<mailto:rti_help+u...@googlegroups.com>.
To post to this group, send email to rti_...@googlegroups.com<mailto:rti_...@googlegroups.com>.
For more options, visit https://groups.google.com/groups/opt_out.

[Orah Maggen]

HI ,

At the 14th seminar on the care and conservation of manuscripts October 2012 Copenhagen I gvae the following presentation " The Illuminated Pixel Adaptation of Infrared and Visible Light RTI in Exploration of Medieval Manuscripts.

the paper will be publish next year in the Conference proceedings.

 

Michael Maggen Head of Paper Conservation Laboratory the Israel Museum Jerusalem Israel

2013/3/23 Earl G.P. <graem...@soton.ac.uk>

To unsubscribe from this group and stop receiving emails from it, send an email to rti_help+u...@googlegroups.com.
To post to this group, send email to rti_...@googlegroups.com.

[Taylor Bennett]

Graeme, thanks for your interest; we'll see!  The poster/demonstration is intended as a proof-of-concept for a combination of open-source methods for image analysis (multispectral RTI, Meshlab, ImageJ with DStretch plugin, etc.), and also for the portable set-up to control the camera motion for detailed imaging of a fairly large object surface.  If the results are interesting enough, I might look to publish them in the future.  So far, there are some intriguing partial fingerprints (one in the paint and a clearer one on the stretcher), and it remains to be determined if they can be identified.  I'm looking for exemplars with which to compare them, and have corresponded with one museum toward this end.  Eventually, we hope that an expert will be interested to take a look at the painting and tell us whether additional tests are warranted that would help shed light on the painting's origins.  If any of this group's members will be attending ICCP 2013, I'll look forward to seeing you there.

Best,
Taylor

On Saturday, March 23, 2013 8:44:06 AM UTC-7, Graeme Earl wrote:

Sounds really exciting! When will it be published?

All the best. Graeme

On 20 Mar 2013, at 09:44, "Taylor" <taylorb...@gmail.com<mailto:taylorb...@gmail.com>> wrote:

A quick update:  I'll be presenting a poster/demonstration of my RTI project at the International Conference on Computational Photography (ICCP 2013) which is being held at Harvard University on April 19-21, 2013.  The poster/demonstration is titled "High-Resolution Multispectral Reflectance Transformation Imaging of a 19th Century British Painting."  The poster/demonstrations are presented from 8:30 to 10:30 am on Sunday, April 21.  For more information about the conference, you can visit http://www.iccp13.org/

Best,
Taylor


On Saturday, January 5, 2013 1:12:11 PM UTC-8, Kathryn Piquette wrote:
Happy New Year, all! Does anyone have any recently published or forthcoming work dealing with multi-spectral RTI to report?

Cheers - Kathryn

--
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 unsubscribe from this group and stop receiving emails from it, send an email to rti_help+u...@googlegroups.com<mailto:rti_help+unsubscribe@googlegroups.com>.

[Earl G.P.]

That sounds fascinating! Good luck with the paper. All the best,

 

Graeme

 

From: rti_...@googlegroups.com [mailto:rti_...@googlegroups.com] On Behalf Of Taylor Bennett
Sent: 24 March 2013 08:26
To: rti_...@googlegroups.com
Subject: Re: [RTI] Re: Multi-spectral RTI

 

Graeme, thanks for your interest; we'll see!  The poster/demonstration is intended as a proof-of-concept for a combination of open-source methods for image analysis (multispectral RTI, Meshlab, ImageJ with DStretch plugin, etc.), and also for the portable set-up to control the camera motion for detailed imaging of a fairly large object surface.  If the results are interesting enough, I might look to publish them in the future.  So far, there are some intriguing partial fingerprints (one in the paint and a clearer one on the stretcher), and it remains to be determined if they can be identified.  I'm looking for exemplars with which to compare them, and have corresponded with one museum toward this end.  Eventually, we hope that an expert will be interested to take a look at the painting and tell us whether additional tests are warranted that would help shed light on the painting's origins.  If any of this group's members will be attending ICCP 2013, I'll look forward to seeing you there.

To unsubscribe from this group and stop receiving emails from it, send an email to rti_help+u...@googlegroups.com<mailto:rti_help+u...@googlegroups.com>.

To post to this group, send email to rti_...@googlegroups.com<mailto:rti_...@googlegroups.com>.
For more options, visit https://groups.google.com/groups/opt_out.

--
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 unsubscribe from this group and stop receiving emails from it, send an email to rti_help+u...@googlegroups.com.
To post to this group, send email to rti_...@googlegroups.com.