Dorin Boyd remembers the first time she saw a hint of slavery from outer space. A satellite image from 2017 in Rajasthan, India, showed a brown oval, which looked like a high-grade high school track. But this was not so harmless: she knew it was a brick oven, one of the tens of thousands of people in South Asia, who often work on forced labor. Boyd, director of the data program at the Rights Lab at the University of Nottingham in the United Kingdom, realized that such images could help her pick up the stove, allowing locally-based organizations to target slaveowners on sites. "You can not see slavery directly, but you can conclude this," she says.
The growth of the number of Earth observation satellites, as well as the improvement of algorithms that can interpret the flow of data they provide, set the goal. modern slavery under the spotlight. This week, at a conference in New York organized by the University of the United Nations (UNU), computer experts, slavery experts and strategy strategists, they presented their latest efforts in their fields of work and methods of collaborative work that took place by storm. . "We do teamwork," says Austin Choi-Fitzpatrick, an expert on peace research at the University of San Diego, California, who interviewed slave owners on stove sites like those studied by space laboratories. According to the latest estimates of the International Labor Organization, headquartered in Geneva, Switzerland, today people are in captivity. But finding them is difficult. "People affected by this are often concealed from the state's point of view," says James Cockayne, director of the UNU Policy Research Center in New York, who helped organize the conference. Boyd estimates, however, that one-third of all slavery is visible from outer space, whether it is in screams of stoves, in illegal mines, or in the contours of transition camps for fish processing
In 2015, DigitalGlobe, whose Earth observation satellites provide most of the data for Google Earth, recruiting users to zoom in on the images of Lake Volta Ghana, where experts suspect that children are forced to work in fishing. "We watched through this massive lake to try and locate the boats," says Rhiannan Price, director of GlobalGlobe Global Development in Westminster, Colorado. In total, 90,000 users have secured 80,000 boats, buildings and fish cages.
Boyd now uses artificial intelligence to speed up the search. As a pilot project, she and her counterparts at the Laboratory used visual searchers to identify brick kilns. The oval form of large ovens, sometimes 150 meters long, and their chimneys differ even from space. "You can not mix them with something else," says Boyd.
Since then, Boyd has turned to machine learning algorithms that recognize the furnace after they have been trained on examples marked by a person. Last month, in the Remote Sensing magazine, she and her colleagues reported that the algorithms could correctly identify 169 of the 178 furnaces in Google Earth data in one of Rajasthan's areas, although they also detected nine false alarms. , has about 150 small satellites, which daily take pictures of the entire globe of the globe. Images have lower resolution than DigitalGlobe, but their frequency opens up the ability to detect changes over time. "Every day, we see every building, every field, every mine, every car, every forest," says Andrew Zollie, vice president of the planet of global influence initiatives in New York.
plan to investigate high-speed signatures of slavery. From outer space, you can observe the harvesting of cotton in Turkmenistan, and, based on how quickly cotton disappears, you can determine if the cars or hands are picked up. In Sundarban, an area covering India and Bangladesh, shrimp farms and fish processing camps use slave labor to clean mangrove trees – a process that satellites can fix
. The Sentinel-1 satellite of the European Space Agency uses a radar to measure minimal altitude changes, which may reveal the land subsidence of tunnels of mining operations with illegal workers inside. The Sentinel-2 satellite, which detects infrared frequencies, can mark extraction based on the prints of re-open minerals.
Low decisions, even reasonable, are not a means of treatment. Other organizations have to turn detective work from space to action locally. And even before the start of the analysis, researchers should know what they are looking for. "The images themselves are much less worthwhile if you do not know the local knowledge about what you are watching," says Zolli.