"Plan" of the site - essentially an overhead view of the point cloud.
“Plan” of the site – essentially an overhead view of the point cloud.

Working at a research degree, especially in a context like UCL, has the added value of being exposed to some of the latest in research, and in technology. And, in an era where archaeology has moved from the realm of the simple trowel and shovel, it has sometimes become quite an expensive fare. But not always.

The PhD work of a particular colleague of mine came as a pleasant surprise. Susie Green is currently working on the application of structure from motion in archaeology. I will not attempt to explain the technique here, but only briefly its application. Essentially, it is a question of using computer software (which is freeware) to process a series of photographs. Through some weird and wonderful wizardry – which I will leave to the mathematics and computer buffs – the software calculates the various positions from where the pictures where shot, create a point cloud, and place the various pictures in 3D. Using the point cloud, moreover, it is also quite possible to create the plans, sections, etc. for a given site.

I had played around briefly with a few photos and some software back in December. Here’s the door of the cathedral in Rochester, with a Photosynth based on only four pictures. A good start.

Taking the chance of some days in Malta, I decided to give the technique a better try, and use it on a site that I’ve been meaning to document more properly. I took a whole set of photographs (just over 100, to be precise) to document these two spaces from a number of angles, keeping in mind that such a variety will give the software a better chance of getting more details. The result was quite impressive. The image here is the “plan” resulting from the overhead view of the point cloud. And, so far, I’ve only used Photosynth (following the advice in Susie’s blog) to have a vague idea of how promising the result’s might be.

In all fairness, I’m not pretending this to be of the same quality – yet – as a 3D laser scan. But then, it only cost me two hours work to take the pictures, with the only equipment being a ladder, a good friend (in case I fall off the ladder!), my digital camera (my trusty six-year-old Canon 400D) and a four-year old laptop which has seen far better days, and software that is readily and legally available online, making the use of  “structure from motion” readily accessible to all archaeologists. Documenting a site will never be the same, even for mere mortals working on a shoestring budget.

All the best, Susie! And thanks for sharing your work!

Advertisements