The

Volcano

logists

LENS

AMcG It was an educated guess as well.

The architecture of the sensors has

changed a lot in smartphones over the

last couple of years.

On a previous design, the front-

illuminated CMOS, you’d have the active

area, which detects incoming photons,

then you’d have all the wiring at the top.

So the light gets blocked by all the wiring.

But the back-illuminated CMOS, which

was developed by NASA, all the wiring’s

on the back. That technology has only

been applied at the smartphone camera

level relatively recently.

We knew that this would provide far

greater possibilities in terms of getting

UV sensitivity. We knew that people had

done this sort of thing previously, but with

SLR cameras, removing the Bayer filter

and getting UV

sensitivity. To our

knowledge, no-

one had tried it

with smartphone

cameras. But,

based on the

physics and the

architecture

of the sensor,

we had an

idea that this

would be a sensible thing to try. And

if it didn’t work, it didn’t matter; we’d

just spend a few hundred pounds

destroying sensors.

Our end-user market is volcanologists

who live in parts of the world where there

isn’t a whole lot of money, so part of the

motivation is to develop low-cost gear. It

wasn’t just a case of our wanting this for

our research purposes; it was also a case

of making things accessible for others.

We’re now involved in a project where

we’re starting to get the bits of kit out into

the field to end-users. Tom’s just been in

northern Chile with guys who are using

some of our gear. We’re hoping to set

something up in Peru.

HS Can you give the equipment to people

in the field and have them start using

it straight away, or does it take a bit

of training?

TW The more detail you go into with

the cameras and the physics, the more

complicated retrievals of SO 2 you can

get, but a lot of volcanologists know a lot

about rocks, but don’t necessarily know

a lot about gas measurements, so they

want to be able to use this technique

without needing to know all this detailed

theory. So the idea is that this can be

relatively simple.

A lot of the commercial systems

don’t come with code anyway, so people

have got to write it. There are so many

volcanologists groups writing their own

code to do the same things. Recently a

paper released some Python code that’s

open source that can do all of this. My

thought is that it’s a great piece of code,

but the problem is it’s still not very useful

because it still

doesn’t have a

GUI attached.

Maybe more

commercial

systems assume

too much

knowledge

from other

volcanologists,

either in

computer

programming or in gas monitoring. We

try to simplify everything so the end-

user can point and click. SO 2 absorbs UV

at a certain wavelength. So we take two

images: one wavelength where the SO 2

absorbs, then we do another wavelength

where it doesn’t, and we basically contrast

them. We essentially divide one image by

the other.

HS Apart from Raspberry Pi hardware,

how else are you keeping the cost down?

TW The lens caps and mountings are

3D-printed, it’s relatively simple, and

there’s Blu Tack used at the moment to

help things fit together. This is graphite-

reinforced nylon, printed using selective

laser sintering. When I first tried, I did

try with typical fused filament, but found

that, because of the thread on it, I couldn’t

print with the detail we needed. There’s

”

”

There are so many

volcanologists

groups writing

their own

code to do the

same things