The

Volcano

logists

The Volcanologists

INTERVIEW

it. We couldn’t get the grant that we

asked for, so we decided that we’d make

one ourselves.

We started looking at hacking the

Raspberry Pi Camera Module, and it’s

surprisingly sensitive to UV, once you

take the Bayer layer off. There are certain

ways that you do that to make it as good

as possible, and it’s actually better than

some systems you can buy for £10K.

Tom Wilkes I guess the idea came

from astronomy, but it’s actually just a

smartphone sensor so it’s modified gear

that’s actually just used for basically

anything you could think of. People use

it for surveillance, wildlife studies, so it’s

just your bog-standard camera sensor

that you’d find in a smartphone. What’s

great about it is how cheap it is; that’s

the main thing. Because so many people

use these cameras, they have such wide

applications, they can be made for next

to nothing, and so that’s the advantage

of this.

You can use either type of Raspberry Pi

Camera Module; once we remove the lens

system, they’re exactly the same sensor

(the NoIR just has the near-infrared filter

removed from the lens system).

AMcG The key issue is that we strip

the sensor down so that it can see

ultraviolet light, and then we rebuild

the system with new lenses that are UV

transmissive. There’s a thing called the

Bayer filter on the sensor that makes

it see red, green, and blue pixels to get

colour images, but for what we do, we

need to see ultraviolet light, and that

”

”

Using the

Raspberry Pi

Camera Module is

actually better than

some the systems

you can buy for £10K

HackSpace You’re monitoring volcanic

activity, with a Raspberry Pi Camera

Module. How on earth does that work?

We would have thought you’d be taking

samples of gases, or using a seismograph

to measure vibrations under the earth.

How does taking photographs help you

discover what’s going on in the centre of

a volcano?

Andrew McGonigle One of the main

components of volcanic gas release is

sulphur dioxide [SO 2 ], which absorbs

really strongly in the UV range at a

wavelength of about 300 nanometres. The

gases come out of the top, so if you can

develop a UV imaging camera and point

it at the volcano, some pixels are going to

be darker, because of that absorption of

the UV light. You can see that absorption,

and you can process that to work

out the amount of gas coming out

per second.

What you end up with is

something that looks like the output

of a thermal camera, but showing

gases escaping, rather than heat.

Scientific-grade UV cameras are

really expensive. It was almost an

accident that we started working

with Raspberry Pi technology, but a

good accident, in terms of costs.

HS How expensive is ‘really expensive’?

Tom Pering The UV cameras traditionally

used in volcanology are meant for

things like astronomy, and they’re

adapted for our purposes: that’s why

they’re so expensive. The commercial

systems that we used to use, and

that our colleagues in Italy use, cost

somewhere in the region of £10K to

£14K, that level. Using Raspberry Pi, we

can make one for a couple of hundred

quid, which is absolutely amazing. Even

the design and development stage, we

probably only spent four or five thousand

pounds. We undercut the commercial

system immediately.

Part of the reason we started looking

into this is that we couldn’t buy the

commercial system: we couldn’t afford

filter absorbs all the UV. What we did

was we worked out – after a lot of trial

and error – a way of getting that filter off

without killing the sensor.

It just so happened that the Raspberry

Pi version 1 camera had a sensor where

we could, relatively easily, get it off. The

version 2 cameras, we think we’ve done

it with hydrofluoric acid, but that’s not

really ideal. For which reason, when the

version 1 cameras were discontinued, we

bought in 400 of them, which are sitting

in my office, so we’re future-proofed.

HS And what version of the Raspberry

Pi are you using at the moment?

AMcG We’re using Raspberry Pi 3 at the

moment, because we need the built-in

wireless capabilities. We connect from

a laptop to the camera using the

wireless that’s built-in. We don’t

need to move to Raspberry Pi 4; I

don’t think this application needs

[the extra power].

HS What was the eureka moment

that led you to build a camera?

TP It was just a random idea.

We needed the cameras, but we

couldn’t afford them. I’d done

some work previously on low-cost

measurements; Tom came along and

gave it a go, and it turned out to work

really well. We tested it at Drax power

station [near Selby] to begin with, then

went to Etna to test it, alongside the

commercial system.

TW We’ve got two versions: one in the

bigger box that uses a USB power pack

to power both Raspberry Pis, and there’s

this one that runs off mobile phone

batteries. I had this running for six/seven

hours off 3.7 V batteries. That’s another

advantage that we don’t think about so

much: the fact that these things use a lot

less power than expensive UV cameras.

That’s really important when you’re

going somewhere as remote as Papua

New Guinea, where power’s an issue.

A 20 amp power bank will last all day.