Forking Chameleon Vision

A number of the core Chameleon Vision contributors will be forking the GPL-licensed project and moving ahead with (almost) a complete rewrite. We will be operating under a completely different name and brand. This fork stemmed from a disagreement about the governance of the Chameleon Vision project, and a desire for more community control and open decision making. We have already written quite a bit of code for the fork, including innovative features like GPU-accelerated vision processing. More details to be announced.

We’re looking for contributors, and we’d love for you to get involved! There are opportunities to write docs, and frontend (Vue.js) and backend (Java) code. PM @bankst or @thatmattguy on CD to get involved!

Now for the fun part: what should we name the project? Post suggestions in the thread!


Lizard Sight





Chameleon Vision 2


I’m sold on this one (I’m a sucker for a good alliterative name!)

Viper Vision

Beacuse this is an open source project, and it’s a fork of Chameleon Vision that you don’t want to intrude on the branding of, I propose the name “OpenCV”, which will surely not conflict with other projects.

On a more serious note, maybe “Chameleon ReVision”?


Chameleon Vision 2: Electric Boogaloo


Maybe you can partner with @Ryan_Dognaux to create Thrifty Vision…



If you can get a PCB with the LEDs, camera, and power management on it, and just have it plug directly onto a Raspberry PI’s GPIO pins, if the total cost beat the Limelight by a significant amount, we would definitely consider switching over.

As a bonus, this wouldn’t need to be restricted to any software set if you made public how to control the LEDs or access the camera.

1 Like

Pretty sure one of the (now former) Chameleon Vision devs was working on that. No idea if it’s still happening anymore, since I last heard they had production issues.

GPIO and PWM support is definitely something that is on our roadmap. This includes allowing the user to specify python/bash command/scripts to execute to allow them to control pins on the GPIO header along with using 3rd party libraries such as Pi4J. If there is anything specific you’d like to see, I would love to hear it!


It would not be that difficult to control GPIO and PWN with pigpio using shell commands executed through Java.

1 Like

Communicating with the socket interface like the Python lib does would probably be the more proper way to do that. Or rewrite in C++ instead of Java and use the C lib :stuck_out_tongue:

There was oscareye which is kinda dead kinda not right now, which was a PCB with LEDs and power and a enclosure with some assembly needed to a Pi.

Java was chosen for the extreme ease of distribution and cross-platform compatibility compared to C++. The ability to use the pre-packaged and maintained WPILib cscore and opencv builds is impossible to pass up in my opinion.
As Xzibit mentioned, GPIO and other hardware dependent systems are in the works, and likely to be implemented with shell commands as Caleb mentioned.


That was tongue-in-cheek, I can totally understand why you would want to stick with Java for the rewrite :+1:

Glad to hear w.r.t. GPIO support! That would be awesome.

1 Like

Raise the Vision