Long range US X-Ray vision system sees threats from a mile away

According to the National Institute of Standards and Technology (NIST), the U.S. government spends billions of dollars on X-Ray imaging technology for security purposes. The latest project in this realm involves a new class of X-ray tools to “see” things from a kilometer away - just over half a mile - even with incomplete or messy data.

In a press release issued on Wednesday, RTX BBN Technologies announced that it had been contracted by the Defense Advanced Research Project Agency (DARPA) to develop long-range X-ray imaging algorithms. Fascinatingly, BBN’s approach primarily relies on simulations based on a large number of low-quality samples, as opposed to a handful of high-resolution images.

The goal is to create a mathematical algorithm capable of working with relatively shaky or fragmented images, which is closer to what commanders would actually come across in the battlefield, the company said in the release.

BBN Technologies, based in Massachusetts, is a subsidiary of RTX, a Virginia-based aerospace and defense manufacturer. Since its founding in 1948, BBN has contributed some major breakthroughs to networking technologies, a good number of which were also backed by DARPA.

The new initiative, on the other hand, is part of DARPA’s X-ray Extreme-range Non-imaging Analysis (XENA) program. XENA’s aim is to develop “algorithmic toolsets” to infer the interior composition of faraway objects. Specifically, the program seeks to address how motion blur often lowers the quality of intelligence-related data in military settings.

For context, domestic and military law enforcement officials have long used X-rays to “find concealed objects, including threats and contraband,” according to NIST. But these technologies have typically been limited to short distances, as motion blur and noise corrupt the transmission of data the farther the device travels from the target.

As a result, a major challenge of developing advanced imaging systems has been in collecting large, high-quality data to train the system for real-world applications.

BBN’s approach turns this limitation on its head, opting instead to work with the low-quality - yet easier to collect in abundance - data to build an algorithm that’s capable of addressing practical problems out in the field. Importantly, this would still be the case at distances up to about 0.6 miles (1 kilometer).

“We are developing algorithms that turn a small number of grainy snapshots into enough detail for decision-makers to act, whether the mission is assessing potential threats or supporting emergency response operations,” Joshua Fasching, BBN’s principal investigator for the project, said in the statement.

Assuming all goes according to plan, the new algorithm will grant service members “access to actionable information about concealed threats, potential weapons, or structural vulnerabilities from ranges previously beyond reach,” the company said.