Deployable computer vision software that gives drone platforms and detection systems the autonomy to see, navigate and act. Continuously shaped by operational feedback from the most demanding environments.
Engineering Philosophy
We build autonomous systems on precise mathematical thinking, not AI branding. We make perception work in the real world.
Our foundation is 3D vision, spatial geometry, and classical computer vision. Computer vision gives us a direct, geometric understanding of the physical world. That means operators work with systems whose behaviour they can anticipate, verify, and trust under pressure. We use machine learning where it measurably improves an outcome, and everything else stays explainable, verifiable, and robust. Every capability is validated in active operational environments before we call it ready.
Intercept Capability
Onboard autonomy software that closes the loop from detection to terminal guidance. A threat enters the airspace, is detected and tracked, and the intercepting platform executes terminal guidance.
Core Capabilities
Distributed Detection · Visual/Thermal
01 — CHAIN HOME
A distributed early-warning detection capability using low-cost visual and thermal sensors paired with 3D computer vision. Designed to be easily replaceable, creating an asymmetric cost per sensor for the other side. Nodes run standalone or as a network, scaling from single-point protection to wide-area coverage across complex terrain and dense urban environments. Multiple overlapping perspectives reduce blind spots against small drones hiding in clutter, flying low, or approaching along unexpected vectors.
GNSS-Independent · Visual Navigation
02 — PATHFINDER
Pathfinder gives aerial platforms accurate positioning using visual odometry and terrain recognition. No GNSS signal required. Unlike downward-looking approaches, Pathfinder uses sideways-facing vision, which is faster to match and more reliable across varying altitudes and terrain types. The system builds its own position from what it sees and how it moves, and stays stable over time, even in unfamiliar terrain.
Air-to-Air · Terminal Guidance
03 — INTERCEPT
Onboard autonomy software that gives drone platforms the capability to detect, track, and perform terminal guidance against airborne threats. Long-range camera-based detection enables early target acquisition. Intercept runs at the platform level, with or without C2 infrastructure, and reduces the training required for effective drone interception.
Built for Degraded Environments
Engineered for contested, unpredictable conditions. Validated in the field.
3D Vision & Spatial Geometry
Depth estimation, terrain recognition, and geometric reasoning that deliver superhuman precision no matter the environment.
Tracking & Autonomy at the Edge
Detection, navigation, engagement. Multiple capabilities, one connected system.
AI Only Where It Matters
Machine learning applied selectively. Every critical path stays explainable, verifiable, and reliable. Our precision does not depend on data-rich conditions. Computer vision and geometry generalise to the front line and run on low-cost edge hardware.
See It Working. Request a Demo
Request demoHorizon
We build and validate our technology in the most demanding operational conditions. Reliability is our baseline. What we learn in deployment feeds directly back into the product. Our goal is to carry that front-line discipline into civilian and critical infrastructure domains.
People
Computer vision, aerospace engineering, applied mathematics, and 3D graphics, all under one roof. The breadth is deliberate. It lets us build capabilities that no single discipline could produce, and it means we develop our own expertise rather than borrowing it.
We build systems that earn trust through performance. If that matters to you, we want to hear from you. Send an open application.
See Opportunities →Get Started
Schedule a demonstration with our engineering team. We'll walk through Chain Home, Pathfinder, and Intercept, individually or as an integrated ecosystem, and discuss how they apply to your operational context.
