Researchers from the University of California, Santa Barbara, have developed an artificial intelligence (AI)-powered hyperspectral imaging tool for detecting real-time methane emissions and tracing them to their sources.

The MethaneMapper reportedly processes hyperspectral data collected amid overhead airborne scans of the areas under examination.

Source: University of California, Santa BarbaraSource: University of California, Santa Barbara

Because methane is a colorless, odorless greenhouse gas that is an estimated 25 times more potent than carbon dioxide at trapping heat, efforts to tighten controls over super emitting leaks are being investigated.

As such, the researchers are using survey images from NASA's Jet Propulsion Laboratory, looking first at pictures starting from 400 nm wavelengths, and at intervals up to 2,500 nm. According to the researchers, this range includes the spectral signatures of hydrocarbons, including that of methane. The researchers explained further that each pixel in the picture features a spectrum and represents a spectral band. Machine learning is then tasked with taking on the collected data to distinguish methane from other hydrocarbons that were captured via the imaging process. Not only does this approach let users see the plume’s magnitude, it can also reportedly reveal the plume’s source, the team explained.

Setting MethaneMapper apart from similar technologies, according to its developers, is the diversity and depth of data gathered from different types of terrain that enables the machine learning model to determine the presence of methane against different backdrops — for instance, assorted topographies, foliage and other backgrounds.

"We curated our own data sets, which cover approximately 4,000 emissions sites. We have the dry states of California, Texas and Arizona. But we have the dense vegetation of the state of Virginia too. So it's pretty diverse," the researchers explained.

The researchers suggest that MethaneMapper's performance accuracy stands around roughly 91%.

An article detailing the MethaneMapper, “MethaneMapper: Spectral Absorption aware Hyperspectral Transformer for Methane Detection,” appears in the journal arXiv.

For more information on the MethaneMapper, watch the accompanying video that appears courtesy of the University of California, Santa Barbara.

To contact the author of this article, email