Raytheon demonstrates DARPA FENCE event-based MWIR camera for real-time high-speed target tracking

ARLINGTON, Va. - Raytheon, an RTX business in Arlington, Va., has demonstrated an event-based mid-wave infrared (MWIR) camera designed to track high-speed targets in real time while significantly reducing data processing and power requirements.

During a recent demonstration in Northern California, the event-based MWIR camera tracked multiple targets, including ground vehicles, aircraft, and live fires. By capturing only pixel-level motion instead of full frames, the system delivered a near-instantaneous view of activity and resolved rapid motion that conventional infrared cameras often miss.

The prototype sensor was developed under the Defense Advanced Research Projects Agency’s (DARPA) Fast Event-based Neuromorphic Camera and Electronics (FENCE) program, which is developing a new class of event-driven infrared sensors that process only changes in a scene rather than full image frames.

Related: DARPA seeks high-assurance artificial intelligence proposals under CLARA program

"This technology represents a new way of sensing the world in mid-wave infrared," said Colin Whelan, president of Advanced Technology at Raytheon. "By focusing only on motion instead of recording every frame and analyzing after the fact, we gain the ability to track very fast objects with far less data and processing, enabling much quicker threat detection and response time."

FENCE program

DARPA launched the FENCE program to address limitations of conventional and early event-based imaging systems, particularly their difficulty operating in cluttered, dynamic environments typical of military operations. The program is focused on developing a low-latency, low-power infrared focal plane array paired with embedded processing that can identify relevant spatio-temporal signals at the sensor level.

To achieve this, DARPA selected industry teams led by Raytheon, BAE Systems, and Northrop Grumman to develop key elements of the architecture, including asynchronous read-out integrated circuits and tightly coupled processing layers. These components are designed to operate as an integrated sensor capable of filtering and prioritizing relevant motion while suppressing background clutter, with a target power consumption of less than 1.5 watts.

Related: Military researchers approach industry for innovations in orbiting space nuclear power design

BAE Systems and Northrop Grumman have contributed to the program by developing complementary neuromorphic sensor architectures and processing approaches. Both companies have focused on low-latency read-out integrated circuits and embedded processing that enable event-based infrared focal plane arrays to extract meaningful spatial and temporal information directly at the sensor.

DARPA has advanced the effort beyond initial design phases by awarding follow-on work to Raytheon and Northrop Grumman to further mature event-based infrared technologies and demonstrate performance in more operationally relevant scenarios.

Event-based sensors

Unlike conventional focal plane arrays that generate continuous image streams at fixed frame rates, event-based sensors operate asynchronously, producing data only when a change in intensity or motion is detected at the pixel level. This reduces bandwidth and compute requirements, enabling lower-latency operation and improved performance in size, weight, and power-constrained environments.

This sensor-level processing approach, often described as intelligent sparsity, reduces the volume of data that must be transmitted or processed downstream by identifying mission-relevant events in real time.

Raytheon’s demonstration highlights the potential for event-based infrared sensing to address these challenges. By transmitting only relevant changes in a scene, the system enables faster detection and tracking of high-speed threats, even in complex environments with multiple moving objects.

Potential applications include battlefield awareness, base protection, missile guidance, and airborne or uncrewed surveillance, where rapid response and low-latency sensing are critical.