Monday 28 April 2014

Laser scanning Sir George Staunton Country Park

So the other day I was able to get hold of a FARO Focus X330 and use it to record some archaeology! The FARO Focus is a top-end machine; as the name suggests it can fire lasers over 330 metres away and still receive the returning beam, in a 360 degree arc. It can be used by one person, as it is quite light and is about the size of a small suitcase. It also doesn't take very long for it to capture millions of data points, which provide an accurate reflection of a building/artefact. These can be downloaded and "stitched" together to create a 3-dimensional model (for info on how to create 3-d models and some software to use them, see my previous blog).


So why was I surveying some buildings for archaeology? More precisely, I haven't told you what I was surveying! I surveyed the Gothic Library and the Beacon at the Sir George Staunton Country Park in Havant, near Portsmouth. To those who are not familiar with Sir George Staunton Country Park, it is named after Sir George Staunton, who built up the estate with land purchases over a number of years and created the "pleasure grounds" to house his exotic plants and later his memorials (read on below for more info). The buildings in question are two late Regency-period buildings (both constructed in the 1830's), and the Library was one of the first Regency buildings in Britain to exhibit gothic features. It is octagonal in shape, and has lovely pointy bits on the edges of the roof, plus one very tall chimney on one of these edges (possibly related to the old fireplace inside). It used to be connected to Sir George Staunton's mansion, who was an important British ambassador to China in this period, particularly given the Opium wars of the 1830's which made it even more difficult to acquire Chinese literature. It was illegal for Chinese officials to teach Chinese, or give Chinese writings to foreigners, so it becomes all the more impressive to know that Sir Staunton translated the first law books from Chinese into English, and was one of the founding members of the Royal Asiatic Society! The Library's walls alternate between mortar and brick faces, so I wanted to capture this in detail, to see how they might relate to the old mansion. However, you are also allowed inside the library, and so I wanted to capture any features that may give us evidence that this was used as a library, such as old beams for bookcases, etc. The floor plan of the mansion is very sketchy, and there might have been a second library inside the mansion that was never mentioned...
Wikipedia's image for the gothic library.


The gothic library as a raw scan. Note how some of the bottom is missing, that's because there are some hedges in the way!


This is a shot of the raw scan of the gothic library's interior, and compared to the photo below, it's fairly accurate, capturing all the detail on the walls (except for the windows, which look like a bullet has just gone through it)!



The Beacon, on the other hand, was constructed from the remains of a nearby mansion (Purbrook house), and is entirely coated in mortar, with a brick base. A circular veranda It was built by the same architect who designed the Gothic Library, Lewis Vuillamy and represents a very different side to the architect; the Library was a modern, ahead of it's time structure that preceded the gothic revival, while the Beacon harked back to Ancient Greece and Italy (where he had travelled previously). Because the Beacon is so isolated, it is almost preserved perfectly. But more about that in a second...

Can you tell the virtual laser scan model (bottom) apart from the real thing (top)? This is after the data has been "cleaned up" with Pointools, one of the more common (yet expensive) software programs for point cloud edits!

I only needed to take as many scans as I needed (i.e. as many as would cover the buildings, since the scans will overlap). The roofs were very difficult to capture because they are blocked, but can be seen clearly from Google Earth!. So I ended up with 25 scans for both buildings, taking about 15 minutes each. This took the whole day, but no other technique could capture the data so accurately and precisely! A photograph doesn't give you coordinates like a laser scan does.

So when the data downloaded, it turned out that:

  1. The Gothic Library is structurally in good nick, but a lot of the exterior details didn't come out very well. This is because the hedges and trees got in the way, and there was nothing I could do about this! Only LIDAR penetrates trees and this can only be used on planes. Since you can't fly a plane this close to the ground, then this is the best result I could get in a day. 
  2. But, I was able to analyse the wall thickness, and it turns out there may be some interesting mathematical patterns in the buttresses supporting the library.
  3. The Beacon actually has one or two problems! The edge of the roof has been eroded by natural causes (birds and the weather), and a new hole has been found in the interior, which seems to have been caused by the natural building materials.
There are some other cool things you can do with the data, such as showing you the intensity of the point cloud (or where the laser scanner received the most information from). This can tell you things like how reflective a surface may be, especially if it's in an isolated place where you can't tell what material it's made out of, such as the roof of a cave deep underground.

At the end of the scanning and editing process, the scans can be taken into a different software package like Autodesk's ReCap or AutoCAD and digitised to have walls, windows and so on, adding more realism to the model.


One of the scans showing the intensity of laser points, showing where the laser scan picked up the most points off the walls (from multiple scan). Neat, eh? It looks more solid on one side and on the roof because 3 scans were taken of the gothic library, and these were unevenly distributed around the room.


So you'd be forgiven for thinking that laser scanning is a quick and expensive way of finding cracks in buildings! But it can also reveal new, previously unknown features about a building. The cost is coming down all the time, although at about £25,000, this isn't something that can be done on a whim! As always the right archaeological questions are required.

Saturday 12 April 2014

3D Modelling for Archaeologists

In this blog post, I will show archaeologists what 3D modelling packages are, how they can be used, which ones to use and what I recommend for you!

Disclaimer: these software programs may not be free indefinitely, even if they are free at the time of print. There are many trials of products, such as Photoshop (useful for editing photos, which can be used in modelling), but much of the free software is also constantly being updated/ is freeware, and so may contain bugs, so it may not work properly.


Introduction

3D modelling software does what it says on the tin; it allows you to create 3-dimensional models (if you know autoCAD, you can skip straight to the software section below
). The applications are obvious; building models that can bring a new level of interpretation to your work and allow a new medium to be exploited, so the public can be more informed and enjoy your work much more easily. Modelling software developed out of architecture and mechanical engineering, where the accurate visualization of components that are often not fully visible need to be seen. But in recent years the line ha blurred as to where 3D modelling software stops; it can be seen in many aspects of life, in particular video games. Because of it's flexibility in so many aspects of the modern world, archaeology and heritage-related specific software, or even software that is remotely relevant to us, is difficult to come by (unless you're in the right place at the right time).

Critcisms

A lot of worries and criticisms I have discussed with fellow archaeologists is that there is not enough time to incorporate modelling into many projects, because either the software will not be supported, or the software has to be learned, and this will take time away from the project and actual excavation. These are valid criticisms, but for a project that involves a large and/or  public audience that will find this information online, I argue that a model created using free software will not take very long to learn, the time and financial costs involved are very low, and most computers will be able to handle simple models. Compared to software that you have to pay for (such as AutoCAD), many of these software packages are not very advanced, nor do they provide some of the measuring tools necessary for a detailed reconstruction. But, at a time when the world has been undergoing a social and technological revolution since the creation of the internet, archaeology is being left behind in the technological front. but as a compensation, while we have always taken the hand me downs to begin with, such is the rate of development, in this situation the "hand me down" is free and I recommend it to archaeologists the world over, not least because a lot of these software packages are compatible with each other, especially when it comes to publishing! 

However, one thing that will have to be considered when you build a model is how you actually, well, build it. some complex lines and geometries need to be "thought through" before you start building it up (see fig.1). For example, to build an angular roof that fits a circular building, you will need to think about how to create the height of the roof, and how to model any inconsistencies in the roof. This also leads to the question of just how much detail you want to put into the model (down to the individual brick, for example, or just the walls), which details will be left to the textures (think of textures as coasts of paint for buildings, but can be applied to anything)?


Fig.1: Start thinking outside the box to construct more complex shapes. Made in Sketchup (author's own).

Theory?

Theoretically speaking, where is archaeological modelling? the truth is, I don't think we actually know. Yes, it's dangerous to take archaeological models at face value, but where does the theory come into modelling?  With apologies to David Clarke, we need to "lose our innocence" over modelling. In particular with reconstructing an artefact/building that has been destroyed, there is an element of variability/probability/interpretation that needs to be considered and explained. Is the data reliable enough to allow for multiple reconstructions? Part of this topic is  known as "procedural modelling", and none of the below modelling software can accomodate this. So how do you accomodate differing interpretations?  I think this debate will grow organically as it takes up more space in the archaeological literature.

The Software

So on to the software!

At the beginning of your project (or even the thing that makes you start your project), you see an artefact/object that you think would look nice as a virtual model. This is my personal set of questions/ checklist for what you can model:

  • Can you photograph it from all angles freely (excluding the bottom, but would be nice if you could)?
  • If this isn't possible, can you get hold of more advanced techniques, like laser scanning, instead of photography?
  • Is it appropriate to create a virtual model of the object? If you're only going to use it for presentation, that fine, and these free applications are great, but you may not be asking the right archaeological questions if this is your primary focus. Ask the archaeological questions first, then think about how to model it.
  • Are there more appropriate methods available for what you want to model?
  • If this isn't possible, can you get some surveyed points of the object and hope to photogrammetrise it?
  • Most importantly, can it form the basis of your research question?
Here are the ones I recommend from his list, which vary in usefulness depending on your project/artefact. this is not an exhaustive list by any means, but a full list of free software technologies are on this blog:

  • SketchUp - very good for starting out on modelling in general, but doesn't contain many advanced functions in the free version. Because it is so simple, it is very easy to model large areas quickly, without using up much disk space. You can also add some textures, so a very convincing model can be built up quickly. It is better for modelling buildings, rather than smaller artefacts. It also doesn't allow you to import photos, but it can import snapshots of, and georeference to, Google Earth! as a results, it is very tempting to show excavations or known positions of buried features alongside standing features. Combined with it's own online depository, this means that it is very easy to find multiple models very easily. The greatest problem is when complex architecture (such as creating a finial, see fig 3 below) or smaller artefacts. There is also very in the way of photogrammetric functions. BUT, my main criticism is what Google will do with the data, particularly once it is on their depository (Google sold SketchUp in 2012 to Trimble, but Google still release a free version). Will your model put on Google Earth without your permission, for example? In this increasingly Big-Brother style world that Google is turning us into, are we unwittingly giving Google the very data it needs to accomplish such a task?
 fig.2: A work in progress of the gothick Library, Sir George Staunton Country Park, Havant. the simplicity of the model allows for rapid creation with limited data; this was the first time I used Sketchup, and it only took me an hour to make the building (author's own)!


Fig.3: Photo of the Beacon, Sir George Staunton Country Park, Havant. The finial is the metallic decoration on top, full of complex curves that could not be done through Sketchup, unless you were able to reach the top of the roof and measure all the survey points (author's own)!
The main modelling software packages I recommend are:
  • FreeCAD- this requires more knowledge to work than Sketchup, because FreeCAD can implement advanced motion simulation capabilities, so you can do some fancy things, like make a moving water wheel. It is not as intuitve as Sketchup though.
  • 123D catch and make- This is a suite of programs made by a large company (AutoDesk), they has an interface similar to autoCAD; i.e. they have a viewcube, the mechanics are reasponably intuitive too (although if you haven't used autoCAD, then SketchUp may feel more intuitive). It works mainly with dwg files made by autoCAD.
The main photogrammetric software packages that I recommend are:
  • Cultural Heritage Imaging have been pioneering unique photogrammetric techniques in the USA, and their free software is called Reflectance Transformative Imaging (aka RTI). As it has been specifically designed with cultural heritage specialists and archaeologists in mind. However, if you have never used photogrammetric techniques before, then it is worth going over to their website and reading their manual, because it has two programmes that you can download. The methods employed are easy to understand, but the steps in the program (RTIviewer) tend to feel a bit dislocated from the process of actually stitching the photos! Don't worry, this happens, and if you follow the instructions in their manual, you will get a results. This is probably the best photogrammetric software; however, it is highly recommended to read their website first and then employ their methodology (instead of their capture kit, you can get similar results, using a shiny ball, string and  a tripod for most artefacts/ faces). So this software requires the most planning too, but the results are worth it. One example of RTI being practised in the UK is the Saving Your Cemetery or Church in York.
  • 123D catch-autodesk's photogrammetric software. I haven't tried this, so I can't give my opinion!
Finally, to publish your model online for free, there are on or two ways of doing this. Some archaeologists are using sketchfab, which can upload over 25 file types of models (except for .dwg files from AutoCAD). This also exemplifies the multivocality of modelling; event though you have large Trans-National Corporations involved in this market, there are a huge number of very popular software packages, which are largely inter-compatible, allowing archaeologists to do what we want to do, particularly when it comes to publishing models!

So to summarise, if you think that you could model a building, the chances are that there will be a software package for you, that will be free, easy to learn, can be downloaded to your home PC. SketchUp is recommended if you don't mind sharing your models, but other freeware is available, that can do the job just as well.