Difference between revisions of "Optical tracking system for amateur rockets"

Revision as of 12:09, 1 May 2009

The TRoll French amateur rocketry association is developping KEOPS an optical tracking system to pursue its and others amateur rockets.

All of our software, hardware, and electronics designs are/will be licensed under the GNU GPLv3 and are available from our subversion server. Access is granted on request by sending me an email at yXanXn_XgoXuy@yahoo.fr (suppress all the X).

Hardware

The hardware will be simple but robust as the system has to be deployed and manipulated in harsh environment. It is mainly a 2-axis system with a equipment casing at the bottom holding the electronics.

Keops CAO Screenshots on the side

Electronic

As in our rockets, we'll use a microcontroler from ATMEL, we are currently discussing to obtain samples of their brand new XMega but a recovery solution is the ATMega32. Driving 2 motors via H-bridges and coupled to a W5100 for communication, it is in charge of controling the position and/or the speed of the Elphel camera.

Software

The software is composed of 2 parts :

• the embedded part is coded in C using the AVR port of gcc.
• on the PC side, the code is in python using the openCV library.

Principle

The camera films the rocket and the sky beyond it. A picture is retreived and the rocket is extracted. The tracking algorithm ensures to keep the rocket centered in the picture. If not, commands are sent to the microcontroler to adjust the camera position.

Schedule

• First of all, take control over the camera!

That means to acquire pictures with as low latency as possible and to configure the image size on the fly from the python script.

• Second, acquire some test case pictures and videos.

The creation of the first pictures will be probably gives some quite funny "making off" videos (we'll post them) if we respect what we have intended. The videos will be done during the "Festiciels PACA" regional launch campaign on the 28th and 29th of May, 2009 at Eyguière, near Salon-de-Provence. We have agreements to position the camera in the take-off field to be as representative as the intended situation : the national launch campaign.

• Check the algorithm capability and execution time on one side but also build the 2-axis system and the electronics.

• Final integration and tests

If everything runs without any problem (absolutely any problem) we could be ready this year national campaign, end of August. If not, we'll have a full year to optimize and to higher integrate the whole system.

'Live' history

Today, the 16th of April 2009, I received the camera. Here is the very first picture I took :

Yes, it's my rocket TRoy!

and then immediately after a picture of the Mediterranean Sea because we are from the Côte d'Azur