Industrial machine vision
High-speed inspection, counting and process monitoring past the frame rate a conventional camera can sustain.
Applications
Where this belongs
One sensing core, applied wherever speed, dynamic range or power is the binding constraint. Below are the directions we are actively pursuing — and the ones we would like help aiming.
Defense
Five missions, one sensing core
The common thread: passive, low-SWaP, works through clutter and glare, and runs detection on-board, so it fits contested, bandwidth-limited edge platforms rather than assuming a datalink that will be there. Each badge says how far along that mission is.
Field-demonstrated
Ground point defense
A ground-based detect-and-track sensor for fixed sites and maneuver elements, built to cue the fire control system you already field rather than replace it.
Read more →
Demonstrated on field data
Robotics & navigation
Optical flow, ego-motion removal and visual navigation for robots and uncrewed platforms that move fast, work in poor light, or cannot count on GPS.
Read more →
Demonstrated on field data
Space & missile defense
Event sensing for space domain awareness, from the ground and on orbit. Map the stars, separate the movers, and send tracks instead of pictures.
Read more →
In development
Seekers & guidance
The same module, flying on a low-cost effector. Size, weight, power and unit cost set the design, so the sensor has to be built around them from the start.
Read more →
In development
Small-arms C-UAS
A weapon-mounted sensor that helps a soldier find and hold a small drone, sized for a rifle and powered from what the soldier already carries.
Read more →The economics
Cheap threats need cheap answers
The problem with the current counter-drone posture is arithmetic. Firing an exquisite interceptor at a low-cost drone wins the engagement and loses the exchange, and an adversary willing to buy in volume can keep that trade going indefinitely.
A low-cost interceptor cannot carry a high-cost seeker or a top-end GPU. Whatever rides on it has to be inexpensive, low-power and fast enough to close the loop in the terminal phase. That constraint is the whole design brief for Balor, and it is why the sensing approach matters more than the algorithm.
Why now
Three things moved at once
- The threat arrived. Small uncrewed systems over bases, airfields and infrastructure, with GPS denial routine at the front.
- The technology became fieldable. Sensing and inference now run together on a module small enough to fly, at a power budget that closes.
- The buying changed. Volume procurement of low-cost effectors creates demand for a sensing payload priced to match them.
Commercial
The same core is dual-use
Defense is not the only path to sustainment. Anywhere a conventional camera runs out of frame rate, dynamic range or power headroom, event sensing opens a market.
Maritime safety & autonomy
Collision avoidance for crewed and autonomous vessels in glare, spray and low light.
Infrastructure protection
Passive, always-on drone watch over critical sites, venues and airspace boundaries.
Commercial space awareness
Low-power tracking of objects and debris on orbit from small platforms.
AR / VR & wearables
Ultra-low-power eye, hand and motion tracking inside a headset thermal budget.
Automotive & mobility
High-speed perception for ADAS, robotics and driver monitoring in rapidly changing light.
Market sizing for neuromorphic vision and counter-UAS varies widely between analysts. We treat published figures as order-of-magnitude, not as a forecast.
Which of these should we go after first?
That is a real question, not a rhetorical one. If conventional EO or IR is falling short on one of your programs, or a mission you own is limited by size, power or bandwidth, we would like to hear about it.