Last week TU Forensics’ Tim James and Amber Collings set to work 3D scanning our magnificent crime scene house with our FARO Focus 350s terrestrial laser scanner (pictured below).
For those of you who keep up-to-date with all things Teesside blogging related, you’ll recognise this piece of kit from the latest episode of the TUBA blog detailing a previous FARO scanning adventure to Vindolanda, led by TUBA’s Rhys Williams (check out that blog post here). Being situated on the Middlesbrough campus, scanning of the crime scene house was less of a journey far away and back in time and more of a pop downstairs this time, but as with any data collection this scene bought a fresh new set of challenges and considerations.
Indoor vs Outdoor scanning
While scanning large, outdoor, open sites like an excavation or a field can be daunting due to their vast size alone, other issues crop up too. For example, repetitive landmarks or an altogether lack of landmarks can cause significant problems. Imagine a lush grassy open field – gorgeous of course but distinctly lacking in any landmarks for the scanner to naturally orient itself in space; all four corners of the field look pretty much the same – green, and well, grassy. Now transport yourself to a dense wooded area, now we have lots of trees to orient ourselves with, right? Well not so much, unless each of those trees is entirely distinct from the next. This is less of a problem for the scanner in terms of actually imaging the scene, but more of a problem for the poor individual trying to register all the different scans together! It’s going to very quickly resemble a particularly cryptic 3D jigsaw puzzle with no picture on the box, just a faint memory of the site.
So how do we get around that? It’s pretty straightforward actually, we use targets. Think chequerboards stuck to trees, or white spheres dotted about at various positions. The use of targets and/or spheres brings with it a whole new level of scanning strategy but perhaps that is another story for another blog post.
Forensically, you would be forgiven for getting twitchy about sticking target objects all around your preserved scene but thankfully they are pretty unnecessary when it comes to scanning indoor scenes like houses. This comes down to the fact that rooms in house are usually filled with stuff; furniture, electrical items, ornaments, etc. each acting as a landmark for quick and easy feature matching. Great! Automatic registration. No need for manual alignment of 50 scans that all look the same! Just don’t move anything…
And that is where the challenge of indoor scanning arises. While all the ‘stuff’ makes for useful landmarks, it also makes for not-so-useful scanning obstacles. Whereas an empty room could be captured with a single central scan perhaps, tables, beds, desks etc. all get in the way. Unlike an x-ray, the laser scanner cannot ‘see-through’ objects, the signal bounces back of the surface of any object it contacts (hence being known as a surface scanning technique). Furthermore, the scanner itself covers 360 degrees laterally, completely spinning around its vertical axis, but only captures 320 degrees longitudinally overhead, leaving a blind spot underneath the scanner tripod (inside which one small adult size sitting human can fit – how convenient!). With that in mind then, lots of scans need to be taken from lots of different positions/angles/heights around the room to get the full picture. Not a particularly easy feat if you find yourself scanning a small, cramped, cluttered scene.
Scanning multiple views is not only time consuming then, but also requires a tricky bit of strategising and increases scene contamination risk. The more shuffling about the room you do with your tripod and scanner the more you risk interfering with the ‘stuff’ in the scene, as we found scanning our very own scenes.
Inside the crime scene house
Being a converted terrace house, the crime scene house facilities we have on the Middlesbrough campus present a realistic residential scene. There’s a hallway leading to a toilet, bedroom, and study on first floor, and a kitchen, living room, and two further bedrooms on the second floor. To generate our mock scene, Tim expertly scattered different evidence types throughout the rooms, from bloody marks and discarded weapons, to drug paraphernalia and digital devices.
Overall it took us approximately 8-9 hours of scanning, across two days, to capture the entire scene house and the outdoor area in front of it. Our strategy was to treat the scanning very much like one would approach traditional crime scene photography. That is, starting with the wide scale shots, i.e. capturing the outside area in front of, and leading up to, the front door, before methodically scanning each the approach to and contents of each room.
But why were we doing this and what are we hoping to achieve?
This wasn’t just a day (or two) out of the office, we promise! We are hoping to generate a complete 3D model of our scene house, not just because it looks cool, but because it can be a beneficial tool in research and teaching.
All challenges aside, this technology is becoming extremely useful for many different aspects of criminal investigation, from scene investigation to evidence presentation in court. We hope to use our scene house and scanning technology to conduct research into best practice and scene scanning protocol construction.
Furthermore, we aim to generate interactive teaching materials for our students. Allowing them to virtually perform tasks, such as practising their approach to crime scene photography, using scene models to inform forensic strategy development, or even honing their investigative skills by analysing line of sight and entry/exit routes.
So, what happens next?
Well, now that the scanning is done it’s time to register and align the scans in the FARO software. Then comes the cleaning, making sure the data looks a good as possible, before meshing and creating a fly through video.
Keep your eyes peeled for future updates on Project scene house 3D – hope you are as excited to see the finished masterpiece as we are!
Post Written by: Dr Amber Collings