Vision fun

Would anybody want me to make a simple vision program that detects the targets and will give info on them similar to what i did last year with the yellow totes? I know wpilib already has grip, but it does not seem polished and I do not think they are really focusing on it…The solution ideally run on either the driverstation computer or something such as a powerful coprocessor(odroid) maybe the roborio itself…discuss below, just thinking this could actually be VERY useful if you have a shooter…

(example output image)](

That’d be awesome. I’m just starting to learn opencv and vision. I’ll take a look at your code for the totes.

Alright, since somebody wants it I will open source it…I will make a post about it soon

Definitely please do. The vision tracking I made last year was inspired by your project.

We are looking into vision processing this year. We have some new team members who are interested in programming.

We use C++, but I suppose we can translate your java.

It looks like you are using Network Tables, so I guess this runs on the Drive Station laptop in which case presumably we can use it as is on the laptop.

Are there any issues with a java program and a C++ program communicating with each other through Network Tables?


hmm lets see, you might be able to write your own in c++. Esentially it could be run anywhere, from the driverstation (fastest computer you really have acess too) or a coprocessor) Network tables use a standard communication protocol so it will even work with labview.

Basically what Fague said. There is no reason that you can’t easily port it over to C++. While there will be differences in the syntax, the process outlined will be the same.

in fact, there is even more documentation on c++…

your all going to hate me for doing this so easy, but here is a little sneak peek

the code isnt optimized nor does it really give any good details but hey, it identifies the targets

Tower Tracker is live! click here for the link