-
Notifications
You must be signed in to change notification settings - Fork 107
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Pi 3 deploy issues #765
Comments
For one, make sure that you have ntcore installed on the pi, as the one that ships with grip is built for armel, not armhf (making it incompatible). Once you have that installed, say to /usr/local/lib, you need to add -Djava.library.path=/usr/local/lib to the jvm arguments when you deploy. I spent some time last summer getting GRIP to deploy to the pi and am revisiting it now, so if you get stuck on anything feel free to ask. |
@blucoat Could I impose on you for a little help getting GRIP/NT working on a pi? |
WPI has networktables builds for the pi 3 now. Most current version here |
@kjm16216 What I ended up doing was updating the version of javacv which GRIP depends on to a newer one which supports armhf, and building a custom version of GRIP using that. To install NT, I just built from source on the pi, but it looks like now there's a better way to do it. Also, I ran into some strange issue with javacv not detecting the architecture right. It looks like it's been fixed, but I haven't had a chance to test it yet. Before you put too much effort into that, it looks like the GRIP developers really want you to use autogenerated code, rather than deploying now. Our team bought a pi and I really would like to isolate vision code from the robot code, both for performance and safety reasons, but as of right now I'm most concerned with getting something that works quickly. It looks like this year's version of WPILib (at least the Java version, I haven't looked at the others) includes opencv and a big improvement over the old CameraServer. It looks like it's the Java bindings that come with OpenCV, not JavaCV (which is much better IMO), but it's still better than being left to install them yourself. For right now at least, I think we're going to use this approach. Maybe there's a good way to run autogenerated code on the pi, without having to deploy GRIP directly, or maybe with the new changes deploying to the pi will become much easier. Either way, I'd love to see more support for offloading vision code to a pi, but until then, I don't think that's the path I'll be taking. For the next few weeks I'll be mostly dedicated to working on my team's code, and then I'll be going back to school, but after that I might try to seriously work on a better solution for the pi. In the meantime, I can answer any specific questions/problems you run into. |
@blucoat Here's where I'm at: I can open the Raspberry Pi Camera module, capture an image, perform a GRIP generated filter (from autogenerated code), and display the values of the report to the console. Specifically, right now I have code that is supposed to put it out on the network, before the holidays, it would give an error both in Eclipse when I tried to run and on the pi. Then I committed the cardinal sin of changing 2 things at once. |
@kjm16216 To avoid spamming this issue, want to switch to email? Just to make sure, you're talking about a deploy to the RIO from eclipse? It seems that the code can't find the native libraries for wpilib (not even networktables). If you make a new blank project and try to deploy it, does it work? |
Sure, sent! Thanks. |
While trying to delpoy to my Pi 3, I run into the following error(s?):
Based on suggestions made in another isssue, I tried editing the
.bash_profile
file and inserting this:set LD_LIBRARY_PATH=/home/pi/vision/grip:$LD_LIBRARY_PATH
.However whenever I try to edit the file (or search for it) I can not seem to find it. Any suggestions?
The text was updated successfully, but these errors were encountered: