AirfieldWatch AI

Name: AirfieldWatch AI
Brand: IEMI Labs
SKU: airfieldwatch-ai

Thermal and visible bird-hazard detection for regional airfields

Concept · not for sale

Engineering design: AirfieldWatch Survey Post

Target price from$8,900per turret (thermal core options)

Software: $399–$699 per month per turret

I want this Ask a question

Concept, not for sale yet. Tell us if you want it; that decides what we build.

Rate this idea and tell us what to change

AirfieldWatch AI concept: the idea and where each part goes

Drag to rotate · right-drag to move · scroll to zoom · click a part for its size and placeSwipe to rotate · pinch to zoom · two fingers to move · tap a part for its size

How it fits together

AirfieldWatch AI, part by part

Scroll or tap a part. Press “See the 3D CAD design” to watch the model, built from our CAD, light each part up and show how it connects.

Real-time 3D render · drag to rotate · final design may varySwipe sideways to rotate

01
The whole AirfieldWatch AI
A solar mast post watches the airfield day and night. The Pi finds birds and animals and texts wildlife staff over 4G.
02Brain
Raspberry Pi 5 + AI HAT+
The brain: it runs the aw-post program
03
HQ daylight camera
Films the runway at 10 fps in 12 MP
Raspberry Pi 5 + AI HAT+ · CSI-2 ribbon cable · Daylight video
04
16 mm C-mount lens
Sets how far down the runway the camera sees
Raspberry Pi 5 + AI HAT+ · C-mount thread · Light onto the sensor
05
Boson thermal camera
Sees animals and people on the runway at night
Raspberry Pi 5 + AI HAT+ · USB-C (UVC video) · Thermal video
06
AI HAT+ (Hailo-8L)
Runs the bird and animal detector on the device
Raspberry Pi 5 + AI HAT+ · Pi 5 HAT connector · Image tiles and detections
07
4G + GNSS dongle
Sends events and clips, and gives time and position
Raspberry Pi 5 + AI HAT+ · USB · 4G data and GNSS time
08
Solar charge controller
Charges the battery and reports its charge level
Raspberry Pi 5 + AI HAT+ · VE.Direct to USB cable · Battery level, solar watts
09
5 V step-down
Turns the 12 V battery into 5 V for the Pi
Raspberry Pi 5 + AI HAT+ · 5 V power lead · 5 V power
10
All connected
8 key parts, one program, and your phone

What it is

A thermal and visible camera turret on a mast that spots birds and wildlife near runways.

What it does

Detects birds and flocks day and night
Classifies hazard-relevant species groups
Triggers dispersal devices automatically
Logs activity for hazard management plans

Who it's for

Regional, general-aviation, private and military airfields.

The idea

Solar camera posts at the survey stations count birds and wildlife day and night for the airfield's hazard study

The problem today

A wildlife hazard assessment at a small airport is a 12-month study built on monthly dawn and dusk point counts by a biologist, about 100 hours of watching a year. It costs $25k to $100k, and avian radar costs about $133k for a one-year trial, so most of the night and most of the year go unseen.

With AirfieldWatch AI

The biologist leaves 2 to 6 fixed solar posts at the survey stations they already use: the runway-end ponds, the creek fence and the drainage ditch. Each post counts birds by size class and flock, and mammals at night, writes the station tables the report needs and texts wildlife staff when a large flock or a deer is in a watched zone. It advises; people decide what to disperse.

1Post at the stormwater pondCounts gulls and geese loafing on the pond by size and flock; texts staff when a large flock sits there too long
2Post at the west runway endOutside the safety area, looks along the approach end where low-altitude strikes happen; logs birds in the air and on the grass
3Post at the east runway endSame job at the other approach end, so both ends of the runway are covered day and night
4Post on the creek fenceThermal camera at night: a deer or fox inside the fence goes to the patrol as a short clip, and the fence gap goes on the repair list
5Post on the drainage ditchWatches the wet ditch where wading birds feed; feeds the per-station dawn, day, dusk and night tables
6The biologist's review queueLabels species on the day's clips in about 30 minutes a post a month; the report tables build themselves

A small regional airport with one runway, its safety area, a stormwater pond, a creek along the perimeter fence and the terminal, drawn to show where each part goes (not to scale).

How the parts connect

AirfieldWatch AI, wired up

Every part, what links it to the brain, and how the data reaches your phone. A solid line is a cable; a dashed line is a radio link.

Drawn from the AirfieldWatch AI parts list. The 3D view above is built separately, from the CAD.

The program

What controls AirfieldWatch AI

Runs on: Raspberry Pi 5 + AI HAT+
Written in: Node.js with C helpers (language not stated in design)
System: Raspberry Pi OS Lite (64-bit)
AI model: YOLOv8s on the Hailo-8L, run on image tiles

1Capture daylight or thermal frames by sun elevation
2Find moving regions, then detect birds and animals on tiles
3Merge tracks of 1 s or more into events
4Save a 6 s clip and queue the event for upload
5Send the event over MQTT and log battery state

aw-post.tsmain loop

while (true) {
  const mode = sunElevation() > -6 ? 'day' : 'night';
  const frame = mode === 'day' ? vis.capture() : thr.capture();
  const regions = motionMask(frame);        // frame differencing
  const dets = hailo.detect(tiles(frame, regions));  // YOLOv8s
  tracker.update(dets);
  for (const ev of tracker.finished(minSeconds = 1)) {
    ev.zone = zoneOf(ev);                   // polygons drawn per post
    ev.clip = clips.write(ev, 6);           // 6 s clip on the SSD
    mqtt.publish('aw/event', ev, { qos: 1 }); // queued if 4G is down
  }
  power.ladder(veDirect.soc());             // cut work as the battery drops
}

To your phone

How the data reaches AirfieldWatch review portal

Biologist's browser; staff get texts

Survey post
4G · MQTT over TLS
IEMI cloud
Web portal · SMS
Biologist and staff

What you see in the app

One screen for each job AirfieldWatch AI does.

Ground flockGulls settle on the grass by the stormwater pond at survey station S4. After 12 minutes the post texts wildlife staff, who decide whether to haze.
02:37
AirfieldWatch review portalNight mammals
Advisory · Deer at S2 fence02:37
Advisory: 1 deer in thermal at the creek fence, about 140 m out.
Deer
1
Range
140 m
Battery
78%
1. 02:37S2 fence · deer · thermal clip
2. 02:38Advisory text to patrol
3. 02:58Patrol confirms, deer driven out
4. 03:05Fence section on repair list
Night mammalsAfter dark the thermal camera picks up a deer by the creek fence. The patrol gets a 6-second clip, checks it and drives the deer out.
Survey tablesEach month the biologist opens the per-station tables: dawn, day, dusk and night counts, laid out like the hazard assessment report.

App screens with sample data. The interface is in development.

What the app gets

Data AirfieldWatch AI sends to the app
Wildlife event	31 large birds, pond	Each event
Review clip	6 s clip	On review queue
Advisory text	Advisory: ground flock, 12 min	Rule triggered
Battery level	78%	Every 15 min
Coverage status	Day and thermal on	Every 15 min

Outbound-only encrypted link, no open ports. Clips upload only when an event needs review. Alerts are advisory.

CAD design · Design complete (stages 1-6), not built

Designed in code, checked in 3D

12 printed parts, each a single closed solid
Every bought part fitted inside and checked for collisions
Assembled exactly as shown in the 3D CAD view (the button on the picture at the top)

Full engineering design STL files

See the CAD parts render

AirfieldWatch Survey Post printable parts rendered from the CAD

Key specs

AirfieldWatch AI hardware
AI processor	NVIDIA Jetson Orin Nano 8 GB (up to 67 TOPS)
Camera	Thermal core (Teledyne FLIR Boson 640 class; Lepton class on entry model) + 4K visible camera with optical zoom
Sensors	Pan-tilt positioner with absolute encoders · Weather sensor (wind, temperature, rain) · GNSS for time and position
Connectivity	Gigabit Ethernet / fibre · 4G LTE backup · Relay and RS-485 outputs for dispersal devices · REST API and webhooks
Power	100–240 V AC or 24 V DC, 40–60 W with heaters (target)
Enclosure	Grey weatherproof pan-tilt turret with heated windows, on a 3 m frangible mast
Dimensions	Turret 25 × 25 × 35 cm on a 3 m mast
Weight	Turret ~9 kg
Ingress rating	IP66

All target specifications

Detection

Detection target specifications
Large birds	Up to ~800 m (thermal, target)
Flocks	Up to ~1.5 km (target)
Mammals	Up to ~500 m at night (target)
Coverage	360° pan, −30° to +60° tilt

AI and outputs

AI and outputs target specifications
Processor	Jetson Orin Nano 8 GB
Classes	Bird groups, flocks, mammals
Outputs	Alerts, clips, relays, API
Logging	Exportable for WHMP reporting

Power and physical

Power and physical target specifications
Power	AC or 24 V DC, 40–60 W
Turret	25 × 25 × 35 cm
Mast	3 m, frangible option
Operating temp.	−30 °C to 55 °C

Targets for production units; they may change during development.

Pricing

Hardware: $8,900–$14,900 per turret (thermal core options)

Software: $399–$699 per month per turret

Target pricing. Mast foundations, power, network and dispersal devices are quoted per airfield.

Runway End

$8,900 + $399/mo

One turret with Lepton-class thermal for a runway end.

Bird and flock detection
Activity logging
Alerts with clips
Relay for one dispersal device

Airfield

$14,900 + $699/mo

Boson-class thermal and zoom for longer range.

Longer-range thermal
Species-group classification
Multiple dispersal rules
Hazard-plan reports

Network

Custom

Multiple turrets and radar integration.

Multi-turret coverage
Radar fusion (planned)
SLA
Custom reporting

Indicative launch pricing in USD, excluding taxes and shipping. Final pricing is confirmed before any balance is charged.

Community

What do you think of AirfieldWatch AI?

AirfieldWatch AI is in development. Rate the idea, tell us what to change about the features, specs or price, and we'll build it with your feedback.

How much people want it

–

No ratings yet

Follow AirfieldWatch AI

Get build updates and first access when pre-orders open.

Comments

No comments yet. Tell us what you'd want from AirfieldWatch AI.

Questions

Ask about AirfieldWatch AI

Tell us what this would save you, or ask anything. We reply within two business days.

Full name

Company / organisation

Role

Country

Product of interest

Your question or what it would save you

We use your details only to respond to this request. See our privacy policy.

From the product lab · ideation phase

Lab updates for AirfieldWatch AI

All lab updates

28 Oct 2026, 14:20LatestA small airport running a 12-month wildlife hazard assessment needs birds and mammals counted at its survey stations through the whole year, not just at monthly dawn and dusk point counts. A study costs $25k to $100k (USDA quote) and an avian radar trial about $133k (sources in DESIGN.md).
- Drawn idea plan (idea.json): a small regional airport from above with the runway and its safety area, the stormwater pond, the creek fence, the drainage ditch and the terminal; one numbered pin per survey post and one for the biologist review queue
- The product picture is now this drawn plan instead of the plain fallback card; it comes from the design notes, not the CAD
- What we learned again from the design: posts stand outside the runway safety area and only advise; people decide what to disperse
- CAD fix (backlog 3): the compute backplate floated 10 mm off the mast in the 3D view; the bought weatherproof box it screws into is now shown strapped to the mast, so the assembly check passes with no floating parts
- Next: drawn plans for more catalogue products that still show the fallback card
See it in 3D Full research and design

AirfieldWatch AI FAQ

General FAQ

Does AirfieldWatch replace avian radar?

No. Radar covers larger volumes and longer ranges. AirfieldWatch adds camera-based detection, species-group classification and visual evidence at key locations such as runway ends, and gives smaller airfields an affordable starting point. Radar integration is planned.

Is it approved by the FAA?

No. AirfieldWatch is not FAA-certified or approved equipment. It is designed to support your wildlife hazard management programme under Part 139 with data and logs. Siting must follow your airport's rules for objects near runways.

Is video processed on the device?

Yes. Detection and classification run on the turret. Only alerts, short clips and logs are sent to the console. The system is designed to detect wildlife, not to identify people.

How is it installed?

The turret sits on a 3 m mast, with a frangible option, on a small foundation near the runway end, with mains or 24 V power and Ethernet, fibre or LTE. We provide siting guidance during the pilot.

Is AirfieldWatch AI available now?

No. It is a new concept in development. Specifications are target specifications. It is a concept: we are not taking orders, reservations or deposits, and it has no ship date yet.

What warranty will it come with?

We are planning a 2-year limited hardware warranty with advance replacement for subscription customers. Final terms will be published before shipping begins.