Category Archives: On Site Work

April 10 Workshop: the Final Day.

On our last official day of the workshop we returned to a operating two camera, my D300S and Alex’s new D700. ¬†Among our first goals was to finish imaging the weathered side of a stone that we had failed to adequately image due to a poor light distribution.

Here are the results from the Victor Dupont monument:

The decoration is now clearly visible under specular enhancement and can be compared to the better preserved reliefs on the other side. Note particularly the vegetative relief at the top of the monument.

While Alex was at work on the Dupont stone, I took an experienced team to work on a heavily weathered marble monument on the other side of the stream:

Notice that we are no longer using a laptop with the D300s and are instead relying on an external IR remote to trigger the shutter. Here are the results:

While the surface normals were within bounds on the capture, the veinous character of the marble weathering made reading the inscription difficult under specular enhancement. Considerable post processing will be necessary to reveal all the characters.

Alex’s group then moved out towards Sydenham Rd. where we had early found a small immigrant plot. Hitherto we had only imaged English-language monuments but thought it was important to capture the Asian monuments at the cemetery before the end of the project.

In many cases, these small, inexpensive stones were in the most desperate need of recording. They were insecurely fixed in the ground and their low height meant that they were more subject to damage than the larger, elevated monuments elsewhere. Here are the initial results:

Despite the fact that it was virtually impossible to get a good distribution of light positions due to the fact that the stone was so close to the ground, Alex and his crew did a fantastic job. The normal visualization above shows that apart from a small area on the bottom, the surface normals were perfectly captured. Catrina Caira, a MAC student and among our most dedicated RTI students, was on hand to help us with the Chinese. She determined, with some help from a friend, that the date of death of this individual was 1926.

Here was the next stone in the plot:

In this case the script was virtually invisible to the naked eye. Again, Alex and his team did a great job of capturing the normals despite the physical constraints, as the following visualization shows:

Here Catrina again helped us with a transcription. More on these Chinese stones in the future!

My own final capture was a more traditional Cataraqui stone:

Despite some lichen on the monument the specular enhancement was especially impressive:

Some more post-processing will be necessary for a full transcription, but we were very pleased with the results. Again, the surface normal visualization proved an accurate guide to the location of the lichen growth on the monument:

This stone proved to be lots of fun in post processing.

It was definitely gratifying (and humbling) to see our students picking out names and dates from the RTIs much faster than their teachers. A very promising start to RTI at Queen’s University!



Filed under On Site Work, Research, RTI

Results from Saturday April 4, 2010, at the Cataraqui Cemetery

On Easter weekend we headed out with a small, but very enthusiastic group of students to make some new RTI captures and improve on ones we weren’t satisfied with from the last outing. We found our group of five got along well and we worked quickly and carefully together.

Although there was haze on this unseasonally warm day that brought down the overall light intensity, we needed an umbrella to remove the highlights cast by the sun on our two black snooker balls. It was worth it as the RTI Builder had no problem detecting the correct highlights when we processed our images.

The umbrella added a very picturesque element to our candid shots of the captures for that day!

One of our objectives was to redo an unsuccessful capture of the monument for Absalom Day. The marble surface was heavily eroded with the harder metamorphic areas of the marble persisting. The effect was an almost veinous texture that made reading the stone very difficult. Under the Diffuse Gain enhancement the text could be made out reasonably well.

Specular Enhancement is normally the technique we reach for first when trying to transcribe a stone. In the case of Absalom Day’s monument, Specular Enhancement failed to substantially improve readability. Much work remains to be done on standardizing these enhancement techniques…why are some effective when others yield less than ideal results?

One stone that was of particular interest to us was an obelisk-style monument with detailed reliefs on two sides. While the writing was very clear on visual inspection, the fine decoration work could not easily be discerned. Interesting too was the obvious preferential weathering. The North-facing side was little abraded with the mason’s original tool marks clearly visible, while the West-facing side had virtually no tool-marks visible. Surely a good candidate for an RTI study! Topography, however, did not work in our favour.

Here we can see a good light distribution on the sphere for the side commemorating E.E. Johnston. The Surface Normal Visualization clearly shows the pale-blue colour that generally indicates a clean capture. In other words, the surface normals point towards the camera, just what we would expect from a flat surface parallel with the sensor-plane of the camera.

On the side commemorating Victor Dupont, however, we failed to make a clean capture despite the fact that the light distribution on the blend image was not appreciably different. The problem turned out to be simple. There was a rise in the ground from the base of the stone to the outside of our light radius that made taking low shots with the light all but impossible at a distance approximately four times the region of interest. We will attempt another capture with a tighter radius and better light control.

This stone was an obvious candidate for RTI enhancement. We could make next to nothing on the stone and could not even discern a single name. The monument itself is sandstone and therefore very susceptible to weathering and the effects of salt efflorescence. The Specular Enhancement improved matters, but much work remains to be done with both the RTI image and in the archives to make a full transcription.

Leave a comment

Filed under On Site Work, Research, RTI

RTI at the Cataraqui Cemetery March 27, 2010

It was a brisk Saturday afternoon at the end of March when we organized the first official day of the workshop in the cemetery. Many were familiar faces to us from the pilot project, but others were new to RTI. The overcast conditions favoured good RTI captures as the ambient light was relatively low and even.

Notice hats and gloves were necessary for the day’s work!

Some low shots using the monopod. Notice Fernando recording the process (see the end of the post of the fruits of his labour).

If you look carefully at these shots you can see a metal rod protruding from the flash along the line of the string and the cone of light (a Canadian invention, we might add)…more on the “rod technique” in a later post.

Fernando again hard at work recording the process in HD video. While it was a good day for RTI, the light conditions were frustrating for such a seasoned filmmaker.

In order to get as many light positions as possible, we have been making regular use of a step-ladder. Caution is needed, however, on the rolling terrain. As we remind the students, many of the dips in the topography represent long-collapsed coffins. It’s important to remember just where we are…

This is wonderfully clear image the rod extending from the flash to the centre of the object.

Here’s a shot of our final post-processing in the Classics Lounge in Watson Hall. We’re using a brand-new Dell M6500 Mobile Precision Workstation with a 17″ monitor and 8gb of RAM. With this, the process of fitting and manipulating Polynomial Texture Maps is a breeze. With the addition of our Intel Solid State Disk we’ve really broken the sound barrier on the assembly of large RTIs (>100mb). Notice also Sarah and Craig sporting the Queen’s colours, just in case you didn’t know this was a Queen’s University Project!

Here’s great video we put together using Fernando’s footage. Many thanks to Sara Gabova for providing the pleasant voice-over.

Thanks Fernando!

1 Comment

Filed under On Site Work, Research, RTI