Open Source FPV Project - OpenFPV - Blogs

I started an open source project to make FPV with the raspberry pi easy to use for every one and wanna invite developers to join. At the moment it is not available for download.

Current Features:

Recording
Web interface
Low-Latency H264 Streaming (≈120ms)
RESTful API
Customizable
Extendable
Installer (in progress)
Minimal bettery consumption

Roadmap:

Invite more people to join the development team
Complete the installer
More field tests with different setups
Create desktop applications for mac/win with HUD support
Add OculusVR support

If you are interested in the project please let me know, if you want to contribute to the project please contact me: contact@openfpv.org

The project website is http://openfpv.org

I will post updates about the project status in the future.

How the system embedded control panel looks like in the moment:

Thanks to every one who providing blog posts around this topic, especially

Philip from diydrones.com with his great article: http://diydrones.com/xn/detail/705844%3ABlogPost%3A1433488

who gave me the initial input to start it. Philip I invite you to join OpenFPV.

I use Ubiquiti modules for these tests with a circular patch antenna on the ground (luxul) and a CP SPW on the plane with fixed bitrates on 802.11n (19Mbps). I also use a custom coded gstreamer pipeline for decoding and one of the important things to play around with on that is the jitterbuffer, that's basically a budget to deal with retransmissions. So in difficult RF environments or longer distances you probably need to increase this slightly. In the camera I tweaked the GOP in such a way that it would have a low image reacquisition if the stream is lost at the cost of higher bandwidth. In my camera the constant bitrate solution didn't work so well. I reckon that better solutions could reduce frame rate when there's a lot of variation in the image to keep bandwidth in reasonable constraints.

I also used a pandaboard once on the ground to do the decoding, but I wasn't happy at all with that. The hardware chip introduced a 500ms delay, so I ditched that in the end. Also, eventually it's very likely you want to play around with virtual reality overlays over the image, since over wifi you can also relay the telemetry from an APM for example. Then you can do some really neat stuff, because the amount of information at the ground is usually a lot larger.

You need to be a member of diydrones to add comments!

Join diydrones

Comments

tshado June 29, 2014 at 8:36am

@Harmen de Jong

Do you remenber the title of the The original topic starter (which had HD FPV streaming all set up and working)?

If not do you remember who was the author and which solutions where used ?

Kind regards,

JC
Tilman Griesel May 5, 2014 at 3:12am

Thank you Randy, sounds really good! I try to get my hands on a ODroid board.
Developer

Randy May 4, 2014 at 5:50am

Looks like a great project.

Some people on the ardupilot dev team are using the ODroid + Logitech C920 camera. In some experiments that were done by Tridge he found that the ODroid was 10x ~ 20x faster than the RPi for OpenCV work.

Thanks very much to Stephen Gloor for the link in the comments about how to remove the bracket from the Logitech C920 camera!
Stephen Gloor May 3, 2014 at 10:15pm

Also sorry Tilman - did not realise we were hi-jacking the thread.

Will message you if we manage to do anything.
Stephen Gloor May 3, 2014 at 9:48pm

Where did you take off from exactly? I may be able to get a team member to do a sweep tomorrow with our OBC plane. No promises as i am not the one that would do it.
Stephen Gloor May 3, 2014 at 9:28pm

Approximately where did you lose it? If it is accessible we might be able to fly some mapping missions and use Tridge's recognition software to find it.
Brian Greeson May 3, 2014 at 8:54am

Very interesting.. I can't wait to see the results of this project
Gerard Toonstra May 3, 2014 at 6:29am

I use Ubiquiti modules for these tests with a circular patch antenna on the ground (luxul) and a CP SPW on the plane with fixed bitrates on 802.11n (19Mbps). I also use a custom coded gstreamer pipeline for decoding and one of the important things to play around with on that is the jitterbuffer, that's basically a budget to deal with retransmissions. So in difficult RF environments or longer distances you probably need to increase this slightly. In the camera I tweaked the GOP in such a way that it would have a low image reacquisition if the stream is lost at the cost of higher bandwidth. In my camera the constant bitrate solution didn't work so well. I reckon that better solutions could reduce frame rate when there's a lot of variation in the image to keep bandwidth in reasonable constraints.

I also used a pandaboard once on the ground to do the decoding, but I wasn't happy at all with that. The hardware chip introduced a 500ms delay, so I ditched that in the end. Also, eventually it's very likely you want to play around with virtual reality overlays over the image, since over wifi you can also relay the telemetry from an APM for example. Then you can do some really neat stuff, because the amount of information at the ground is usually a lot larger.
Tilman Griesel May 3, 2014 at 4:44am

@Marius van Rijnsoever Great! You are welcome please write me at contact@openfpv.org. I use two ALFA AWUS036NH at the moment.
Developer

Philip May 3, 2014 at 2:42am

Actually, if you are using high quality commercial wireless links, not cheap crappy laptop wireless, the wireless is not a noticeable problem in the link. The issue, is getting affordable wireless links that are engineered to do this well.

There is an article on this site about flying with wifi, and these latency due to wifi questions have been answered very clearly by Mr Adam Conway, an actual expert if commercial WiFi....
http://diydrones.com/profiles/blogs/the-case-for-wifi-in-drones
Phil

of 2

This reply was deleted.