Hidden beneath the surface of the U.S. lies a silent legacy of nearly two centuries of commercial drilling: hundreds of thousands of forgotten oil and gas wells. Known as undocumented orphaned wells (UOWs), these relics have slipped through the cracks of formal records, with no traceable or financially responsible operators. Out of sight and out of mind, they form a dangerous and overlooked footprint of America’s industrial past.

[Read how AI is helping the mining industry at GlobalSpec]

Improperly sealed wells can leak oil, chemicals, or gases into the environment around it. Uncovering UOWs and tracking methane emissions is dedicated work, and researchers are turning toAerial view showing recent oil fields in California cutting-edge technology to rise to the challenge. Armed with artificial intelligence (AI) paired with drones, laser imaging, sensor arrays, and topographical maps, the modern technology seeks out the buried past.

"While AI is a contemporary and rapidly evolving technology, it should not be exclusively associated with modern data sources," said Fabio Ciulla, a postdoctoral fellow at the U.S. Department of Energy's Lawrence Berkeley National Laboratory and lead author of the study on using AI to find UOWs published in Environmental Science & Technology. "AI can enhance our understanding of the past by extracting information from historical data on a scale that was unattainable just a few years ago. The more we go into the future, the more you can also use the past."

Ciulla dove into variety of maps of every size, meticulously detailing chunks of latitude and longitude. The unique secret that each map shared was the consistent symbol for wells: a hollow black circle.

Since 2011, the U.S. Geological Survey has scanned and uploaded almost 200,000 historical topographic maps from between 1884 and 2006.

A multi-layer approach

"For a human being, looking at this circle and recognizing it is extremely easy," Ciulla said. "Until recently, this was the only available method to extract information from these maps, but that strategy does not scale USGS historical topographical map of California circa 1850well if we want to apply it to thousands of maps. This is where artificial intelligence comes into play."

To achieve success, the research crew has to train AI to spot the right symbols buried in a sea of visual noise. The challenge doesn't stop there—the AI needs to perform flawlessly across maps of varying terrains, colors, and from perfectly new, to brittle and discolored.

From the lab to the field, researchers take to the ground and the sky (via drones). They execute careful technologies to literally "strike oil" during the execution of their case study, more of which has been documented here.