How museums are using lidar to stop art theft

Since the recent heist at the Louvre Museum, museum security specialists have been examining technologies designed to detect intruders inside gallery spaces. Among the tools discussed is lidar, which uses pulses of laser light to create a three-dimensional model of a room and detect movement within it.

Martin Vojtek, business director for 3D surveillance at the company Octave, said that the technology can be used to monitor galleries and trigger alerts if someone crosses predefined boundaries near artworks.

Vojtek demonstrated the company’s surveillance software, called Coda Spatial, which Octave markets to airports, hospitals and utility infrastructure. While Octave is one provider in this space, other companies are also exploring lidar for cultural institutions.

A screenshot from Octave's Designer tool showing the visual range of a particular camera model during the security design process. Photo courtesy of Octave

Vojtek said Octave has as many as ten public and private museums in the United States and Europe as customers, with additional institutions considering the technology. Because of non-disclosure agreements with some of them and broader customer confidentiality policies, Vojtek was unable to name them to Urgent Matter.

To create the system, the museum uses drones outside and handheld laser scanners inside to map the building and surrounding property. Those scans are then used to build a 3D twin of the museum. The exterior can also be mapped using drone photography and photogrammetry if the drone lacks lidar capabilities.

Vojtek said the setup process takes about three hours and can be done with a variety of scanners on the market. The company is training its clients on how to complete the setup.

If curators later change the location of art and artifacts in a permanent collection, or a new show is installed, Vojtek said it is an easy process to create new scans. “It's no problem. This is a matter of hours, not like days or months. This is super easy to create a digital twin,” he said.

When the setup is done, lidar sensors installed inside and outside the building detect intruders—including how many there are and the path and speed of their movements—with pinpoint precision, which museum security can then see in real time in the Coda Spatial software.

Vojtek said the digital model is built to scale, allowing security staff to measure distances and movement inside the virtual map as accurately as they would in the real building.

He showed a presentation slide deck that explained the benefit of lidar detection over a standard perimeter video surveillance system. He said lidar can avoid some limitations of traditional cameras, which can be affected by fog, darkness or glare—meaning that intruders may go undetected.

The software also allows security teams to automate camera responses after an intruder is detected. If a museum has installed pan-tilt-zoom, or PTZ, cameras, they can be automatically triggered to track and record intruders once they have been detected by the lidar.

A digital twin created in Octave’s Coda Spatial platform shows a mapped property with lidar sensors and alerts triggered by detected intruders. Photo courtesy of Octave

And museums can set up specific detection zones, say the second-floor window at the Louvre where the thieves broke in, where the software can detect suspicious lingering.

“There can be some special zones around the artifacts, there can be some pre-alarm zones, some loitering zones. We have like eight types of zones, and this is what I think is this is very important for museums to perfectly set up,” Vojtek said.

Some recent museum thefts, like the one at the Louvre, happened while the museum was open and other guests were inside. Vojtek was asked about how the software could detect suspicious lingering if a museum is full of patrons.

“I know it's hard in the museum because the loitering is normal by the artifacts,” he responded. He said the system can detect movement with about 2 centimeters of precision, allowing it to trigger alerts when someone crosses a restricted boundary—like the Louvre’s second-floor window—or reaches too close to an artwork.

Another tool, called Designer, allows planners to test camera and sensor placement on the digital model to identify blind spots before installing equipment. Vojtek demonstrated it by placing a marker representing a physical lidar sensor on the digital model, with a videogame-like thief crouched down behind a wall.

“The purple points are what you can see. And you saw that the guy was blocked by the by the wall. But when I lifted it up, the lidar tilted down and now I can see behind the wall,” he said.

It works with cameras, too. With camera security design, the software allows you to select from hundreds of models available on the market. The software will then respond to how far and how detailed a particular camera model can see, allowing planners to choose a model that best fits their needs.

In the case of a museum heist, this could mean the difference between police being able to quickly identify suspects before items are sold on the black market or melted down.

“So, this is also about how smart you are during the design,” Vojtek said.

Vojtek was asked to provide a ballpark of how much it might cost to set up a lidar surveillance system for a museum with Coda Spatial. While prices are individualized, he said it would cost about $115,000 for the software and everything needed to operate four lidar sensors and four cameras covering a large area.

Security experts like Bill Anderson, the co-founder of Art Guard, have also advocated for increased physical sensors, like the company’s magnetic sensors which detect when an object or picture frame has been jostled, moved or otherwise shifted.

Vojtek said the company’s lidar systems have an open API and are designed to integrate with existing security infrastructure, including physical sensors, alarm systems and access controls, allowing museums to monitor galleries through a combined security platform.

“We are not fighting against cameras. Cameras are our best friends because it's our eyes, but we have to tell them where to look,” he said. “That's why the combination between the lidar and camera is really something.”

Vojtek explained with a demonstration of an example of a digital replica of a prison in Coda Spatial. Inside the software, a guard could click at a particular gate at the prison and a standard video management system for CCTV monitoring could instantly show footage from all nearby cameras.

The feature, Vojtek agreed, could also be useful in monitoring things like protest activity inside of museums to ensure particular artworks aren’t vandalized.

“When we are exporting videos for police, they don't understand how we did it,” Vojtek said, adding that the system helps cut down on police investigation time from having to sift through surveillance video recordings to find applicable footage.

Vojtek said the company is exploring artificial intelligence tools to automate parts of surveillance monitoring, though Octave says human operators will remain involved in the systems.

He said Octave recently reviewed its internal road map and believes it may be able to offer an A.I. agent by the end of 2026, a product it is internally calling Scope.

The company is already using neural networks it calls DeepTection for the classification of subjects detected by the lidar sensors—like whether they’re a human, an animal or a car.

“Now it's the time that, with the combination with cameras, we can ask the camera what is it? Is it guy with a gun? Does he have a backpack?” Vojtek said. “I believe that we can understand better the situation than a human operator.”

Clarifications: An earlier version of this article said Octave could not name its museum clients because of non-disclosure agreements. The article was updated at 12:37 p.m. on March 13, 2026, to clarify that the company cited both non-disclosure agreements and broader customer confidentiality policies as the reason the clients could not be identified. It has also been amended to clarify that Octave said it is exploring artificial intelligence tools to automate parts of surveillance monitoring while keeping human operators involved in the systems..