Skip to content

benjilwhite/CVDrone

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 

Repository files navigation

A DIY Drone Computer Vision Development Platform

image

IMG_1362.1.mov

The goal of this project is to create a relatively inexpensive camera platform that can be mounted on a drone, transmit a 1080p video feed, and remotely control pan, tilt, and optical zoom. An interface to this will make it simple to develop computer vision applications that require camera movements.

Wiring/Communication Diagram

Here's the current layout of the system.

image

Hardware Components

Video Feed

For video I'm using the Walksnail Avatar HD FPV video system. It's designed for low-latency FPV quadcopters, and can transmit a 1080p 60fps video feed back to a base station. The link has favorable qualities for this application, with the manufacturer claiming a range of around 2.5 miles, but some online have been able to achieve greater ranges than that.

Camera Setup

The only issue with this system in this case is that the camera features an extremely wide-angle lens. Such a lens is optimal for flying an FPV drone, but I wanted a more flexible camera that could zoom in/out and still have detail. Ideally the camera could be replaced, but the link between the camera sensor and the video transmitter appears to be a proprietary protocol on some sort of coaxial cable, which I don't have the expertise to reverse-engineer. However, the camera lens is easily removed, and the micro camera version has a 1/2.7" sensor.

The solution was to attach the sensor to a larger lens which would allow for remote zoom and focus. I settled on one of these Raspberry Pi Pan/Tilt/Zoom camera lenses, which are designed for a 1/2.7" sensor, and have servo motors to control zoom and focus from a Raspberry Pi. I used a 3D printed mount to align the sensor and lens in a robust fashion.

Base Station

The Walksnail Avatar video receiver outputs an HDMI signal, which I'm converting into a webcam feed using an Elgato Cam Link. This allows any computer vision application to easily get a video feed and treat it as a normal webcam.

Electronics and Control

image

The electronics setup is currently overcomplicated for what it does, and I'm working on developing a simpler system that reduces the amount of communication required between devices. More to come in the future.

Control Link

To control the camera I'm using an ExpressLRS controller and receiver.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published