# Building an Edge-AI Drone Payload with OAK 4 FFC

This example shows how the OAK 4 FFC can be used as a lightweight drone payload. We mounted it on an off-the-shelf F450
quadcopter, together with two OG05B10 global-shutter camera modules and an off-the-shelf pressure sensor to measure altitude, with
all processing running directly on the drone.

The build is a walkthrough and starting point, not a finished product. The 3D-printed mount is made for the F450, so your mount
may differ, but the wiring stays the same. It is meant as inspiration for your own aerial inspection, mapping, or
autonomous-flight projects.

### Materials used

 * OAK 4 FFC (Kit or barebones)
 * 2× OG05B10 fixed-focus camera modules with the included ribbon cables
 * [SparkFun Pressure Sensor BMP581 (Qwiic)](https://www.sparkfun.com/sparkfun-pressure-sensor-bmp581-qwiic.html)
 * Cable with a Qwiic connector on one end and a 2 mm header on the other
 * DC power cable (included in the OAK 4 FFC Kit)
 * [F450 drone
   kit](https://www.hawks-work.com/products/f450-drone-kit-to-build-diy-450mm-wheelbase-4-axis-multi-rotor-drone-kit-b) with a 3S
   LiPo battery
 * 3D-printed mount and camera module holder
 * M2.5 threaded inserts, M2 and M2.5 standoffs and bolts, 2× 20 mm Allen bolts with lock nuts

## System Overview

 * On-board processing The OAK 4 FFC is powerful enough to run AI right on the drone, so only results need to be transmitted
   instead of raw video
 * Two downward cameras Two OG05B10 global-shutter modules stream at 2592 × 1944, 30 fps
 * Altitude from pressure A BMP581 barometric sensor on the I²C bus gives relative altitude, along with pressure and temperature
 * Direct battery power The OAK 4 FFC accepts 6–32 V, so it runs straight from the drone's 3S battery with no extra regulator
 * Adjustable view The 3D-printed holder lets you tilt the cameras to get the view you need
 * Simple camera setup Camera extrinsics are set in OAK Viewer by entering the cameras' relative positions; for overlapping views,
   dynamic calibration refines them automatically without calibration targets
 * Room to grow Any compatible I²C sensor can be added to the same bus

## Building the App with an AI Coding Agent

With the [Luxonis Agent Skills](https://docs.luxonis.com/software-v3/agent-tools/agent-skills) installed, the coding agent
detected the connected OAK 4 FFC, its two cameras, and the pressure sensor, then built the app from a plain-language prompt:

 1. Streaming Stream both camera feeds to a web interface running on the device.
 2. Telemetry Read the I²C pressure sensor and display pressure, temperature, and relative altitude, with a flight recorder for
    playback.

The app was built and deployed directly to the device in a couple of minutes.

## Typical Use Cases

 * Aerial inspection of infrastructure, roofs, and solar farms
 * Agricultural and terrain surveys
 * Mapping and image stitching from the air
 * A perception module for autonomous and smarter flight

## Coming Soon

> **Blogpost coming soon**
> The full step-by-step blogpost, including the complete bill of materials, wiring details, and 3D-print files, will be linked here soon.
