Sometimes great evils are hidden in plain sight. South Asia’s ‘Brick Belt’, which stretches across swathes of Pakistan, India, Bangladesh and Nepal, has thousands of brick kilns, and many use slave labour.
Now many of these sites have been identified thanks to Doreen Boyd and her expertise in using remote sensors on unmanned aerial vehicles (UAVs), aeroplanes, helicopters or satellites to take observations and measurements of our planet.
Dr Boyd is part of the Rights Lab, a Beacon of Excellence that brings together more than 100 University of Nottingham scholars, including observation scientists, geographers, social scientists, engineers and computer specialists with the aim of helping to eradicate modern slavery. She said: “Satellites are orbiting the Earth all the time. There are currently around 1,300, of which 600 are designed to observe the Earth.”
My work has always been driven by the link between people and pixels.
This profusion of high-resolution, freely available satellite data has enabled Dr Boyd and her team to make the first credible estimate of the number of brick kilns across South Asia and their impact on the environment.
By analysing this data and cross-referring to on-the-ground knowledge from partners, the Rights Lab team is supplying unprecedented insights into where slavery is taking place. This data will allow NGOs and governments to better target the fight against modern slavery and meet the UN’s goal of ending this blight on human history by 2030.
“My work has always been about using data, but the really important aspect is what you use that data for and how you process it.” says Dr Boyd. “Once we had the satellite data, we needed to identify where the brick kilns were – and there was no easy way to do that. My PhD student, Bethany Jackson, had to sit and look at every image to identify what she thought was a brick kiln.
On the ground
“We initially focused on India, as one of our colleagues – Professor Kevin Bales from the School of Politics and International Relations – had been there, and you have to have verification on the ground that what you think you can see is actually there.
“We obviously couldn’t manually look at the whole of the brick belt as it is huge – but by taking a sample area, we were able to come up with a straightforward and easily replicated statistical approach, which allowed us to extrapolate from a random sample.”
Dr Boyd has worked closely with colleagues from the School of Geography (Professor Giles Foody and Dr Jessica Wardlaw) and also with the Faculty of Engineering (Professor Stuart Marsh) on the project. As well as using statistical inference of visual analysis of the satellite imagery, the team are also using Citizen Science, where members of the public identify slavery areas (such as brick kilns) on satellite images and these are verified by cross-referencing to data.
“Now we have the estimate, the next step is to map the location of every single brick kiln using artificial intelligence. We have to ‘teach’ machines to identify the kilns for us,” says Dr Boyd.
Along with colleagues from the School of Computer Science, the team will now train computers using an algorithm to understand what a kiln looks like.
“Once we’ve trained the algorithm, we can be pretty confident that it can look at other images and accurately spot a brick kiln. We can amalgamate this feature identification with the crowd’s results and we can be, say, 95% certain that there are brick kilns in a specific location.”
Once this method has been perfected, the team will pass on the techniques to government agencies and NGOs.
“The work we are doing is all about the Freedom Dividend. If we can free every single slave, we can live in a safer world, a greener world and a more prosperous world and that’s what’s driving our work. My work has always been driven by the link between people and pixels - and looking at what is happening on Earth is something I’ve always done. I’m now excited by the prospect of using the ever-advancing technologies we have to help eradicate modern slavery.”