Drag to Rotate Autonomous · Scroll to Continue
Anor System
Autonomous
Neural
Observation &
Recognition

Autonomous Defense. Vision-Primary.

Anor builds autonomous drone systems that see the world with cameras — not GPS, radar, or RF. The Kestrel Series applies that same vision-primary stack to industrial inspection and counter-drone defense, delivering persistent situational awareness where conventional systems fail.

$20B Market by 2030
0 Radar Dependencies
1 Unified CV Stack
01 — The Problem

1 Interceptor. 20 Drones.
Same Budget.

Counter-drone budgets were sized for single, expensive intercepts against single, expensive threats. That threat model is gone. Adversary drones are mass-produced and disposable — cheap enough to overwhelm any system built to fire one shot at a time.

UNIT_ECONOMICS.LOG
Traditional Effector QTY 1
$1,000,000+
Adversary UAS QTY 20+
<$50,000each
> Budget parity: 1 effector ≈ 20 threats
Figures are illustrative — actual unit costs vary by platform and program.
Legacy Stack
GPS Radar RF Link
Jam any one node, lose the whole system.
Anor Stack
Camera Edge AI Action
Closed loop — nothing external to jam or spoof.
02 — Thesis

The same shift Tesla made over LiDAR —
applied to the aerial internet.

Existing autonomous systems rely on GPS, radar, or RF signals — all jammable, all infrastructure-dependent. Anor builds vision-primary from the ground up. Cameras see. Edge AI decides. The system acts. This is how the aerial internet gets built.

Why It Works

Four structural advantages fall out of building vision-primary from day one — not bolted on, not a fallback mode.

01Unjammable by Design

No RF signal to defeat. No GPS to spoof. Vision systems operate on light. You cannot jam light.

02No Infrastructure

Radar needs towers. GPS needs satellites. Vision needs a camera and power. Deploy anywhere.

03Full-Stack Ownership

Hardware, CV software, flight control, and C2 interface — all designed in-house. No integration risk.

04Platform, Not Product

One CV stack powers Kestrel drones and ground vehicles. Every advancement benefits both.

05LLM Command & Control

Operators issue natural-language directives, not waypoint files. The system translates intent into flight plans and sensor configurations.

06Autonomous Swarm Coordination

One perception stack runs across every node simultaneously — coordinated inspection and defense patterns with no central controller.

03 — Product Platforms

Kestrel Series

One CV stack, forked into two missions. Everything Marr Vision learns on an inspection flight sharpens the interceptor, and vice versa.

The Foundation Marr Vision CV Stack
Scout Class Kestrel Scout Industrial Inspection
Interceptor Class Kestrel Interceptor Counter-Drone Defense
David Marr · MIT · 1945–1980
Proprietary Vision Stack · Anor Industries

Marr
Vision

Anor's proprietary Marr Vision stack layers scene geometry, motion parallax, and semantic understanding into a unified spatial model. The result is a drone that perceives depth, intent, and context — not just bounding boxes.

The Pipeline
01

Sense

Sony IMX477 · RGB Camera

Real-time video. No radar. No lidar. Vision only. The raw material of autonomous intelligence.

02

Detect

YOLOv8 · 30fps · >95% mAP

30fps detection at >95% mAP. Runs on NVIDIA Jetson Orin at 10W. Classify targets, threats, and obstacles in a single forward pass.

03

Track

DeepSORT · Kalman Filter

Multi-frame target locking with Kalman filter prediction. Deep appearance descriptors maintain identity through occlusion.

04

Act

MAVLink · Pixhawk

Vision converts to flight commands. Pixhawk executes in milliseconds. Operators issue natural-language directives through an LLM-powered C2 interface, which translates intent into MAVLink waypoint and sensor commands in real time — no manual flight planning required.

05

Coordinate

Swarm C2 · Distributed

Multiple units share target data. Distributed intelligence. No single point of failure. Swarm-native from day one.

Depth from Parallax · Triangulated Scene Geometry
Named After
“Vision is the process of discovering from images what is present in the world, and where it is.”
David Marr · Vision, 1982 · MIT AI Lab
Level 01 — Computation

What Must Be Seen

Target ID, depth, velocity, trajectory. Define what the system needs to extract from every scene to accomplish the mission.

Level 02 — Algorithm

How It Is Computed

YOLOv8 + DeepSORT + LLM translation. Fast enough for real-time flight decision-making at the edge.

Level 03 — Implementation

The Physical Stack

NVIDIA Jetson Orin, Sony IMX477, Pixhawk. Hardware chosen for field conditions, not lab conditions.

Kestrel Scout System
Anor Industries · Autonomous Systems

Kestrel Series

Two variants. One CV stack. From scanning a pipeline to intercepting a drone swarm. Kestrel sees, decides, and acts without a human in the loop.

KS-01 · Scout Class

Kestrel
Scout

Interact to rotate in full 360°. True to production specifications.

Drag to Rotate • +/− to Zoom Auto-Rotating • +/− to Zoom
Live Simulation

Autonomous Facade Scanning

Kestrel Scout drones survey a building envelope in formation, scanning the structure for cracks, corrosion, and coating anomalies as they fly — relaying findings back to home base in real time.

Industrial Inspection

Kestrel
Scout

An autonomous industrial inspection drone engineered for linear infrastructure — pipelines, power lines, bridges, and solar arrays. Kestrel Scout operates beyond visual line of sight, building a real-time structural health model as it flies. Mission tasking runs through Anor's LLM command interface — operators issue plain-language directives, and multiple units share detections across the same perception stack in real time.

Autonomous waypoint navigation along linear infrastructure
Real-time anomaly detection on pipelines, conductors, and bridges
Sub-millimeter crack and corrosion scoring via IMX477 imagery
Thermal overlay option for heat-signature inspection
Geofenced no-fly zones with automatic avoidance
Mission playback and annotated report export
Technical Specifications
CameraSony IMX477, 12.3 MP
ComputeNVIDIA Jetson Orin NX
Flight ControllerPixhawk 2.8
DetectionYOLOv8 + DeepSORT
AI InterfaceLLM command-and-control
Vision StackProprietary Marr Vision
Max Flight Time62 min
Operating Range15 km (BVLOS)
Wind Tolerance14 m/s
IP RatingIP54
KI-01 · Interceptor Class

Kestrel
Interceptor

Interact to rotate in full 360°. True to production specifications.

Drag to Rotate • +/− to Zoom Auto-Rotating • +/− to Zoom
Live Simulation

Coordinated Threat Neutralization

Kestrel Interceptor units operate across a quantum-resistant mesh network, autonomously identifying and closing on hostile drones — LLM-to-MAVLink command translation drives intercept without a human in the loop.

Counter-Drone Defense

Kestrel
Interceptor

A vision-first counter-drone defense system that acquires, tracks, and intercepts rogue drones without GPS or RF dependence. Kestrel Interceptor is deployed at perimeters, critical sites, and mobile command posts where conventional counter-UAS fails. Engagement is directed through the same LLM command interface, with autonomous swarm coordination scaling a single interceptor into a coordinated multi-node defense grid.

Vision-first target acquisition and lock confirmation
DeepSORT multi-track across swarm densities up to 1,200 objects/km²
Pixhawk-guided intercept with configurable rules of engagement
Counter-GPS and counter-RF operational mode
LLM-generated post-engagement forensic summary
Scalable from single-node to coordinated multi-drone defense
Technical Specifications
CameraDual Sony IMX477, 12.3 MP
ComputeNVIDIA Jetson Orin NX
Flight ControllerPixhawk 2.8
DetectionYOLOv8 + DeepSORT
Tracking CapacityUp to 1,200 objects/km²
AI InterfaceLLM command-and-control
Vision StackProprietary Marr Vision
Max Speed22 m/s intercept
Engagement ModeAutonomous / operator-confirmed
Counter-MeasuresGPS-denied, RF-denied operation
Anor Industries · Get In Touch

Ready to see the
platform in action?

Whether you operate pipelines, power infrastructure, or require counter-drone protection, our team can scope a deployment that fits your environment. Request a demo or reach out with questions.

Request Received

Thanks — we'll get back to you shortly. In the meantime, feel free to explore Marr Vision or the Kestrel Series.