This tutorial elegantly bridges the gap between LEGO’s accessible AI tools and the advanced control of Pybricks, making real-time computer vision practical for any roboticist. It is a significant contribution to democratizing sophisticated machine learning within the STEM community.
Approfondir
Prérequis
- Pas de données disponibles.
Prochaines étapes
- Pas de données disponibles.
Approfondir
AI Hack for LEGO SPIKE Hubs FULL TutorialAjouté :
In this video, I will show you how to install my custom code to let the Lego education coding canvas to stream the AI data to your Lego py powered hubs like spike prime or spike essential. Let's dive in already. So you go on code.leucation.com and you create a new project with wordblocks because there are there is no other option for now. Then you click on post classification just to enable the camera so I can greet you. Hi, below this video you will find a link to my GitHub repository and from there you can click on the readme and you will be brought to this page. Here you find this uh you can find the whole script here but you don't need to read it. You have to trust me and then you go here on you drag this onto your bookmark tab on your um Chrome browser. This works on Chrome and I couldn't find a way to make it work on Bluefi browser on the iPad. So if there are any hackers among you that can help me to contribute to this, you are welcome. I can make it work on Android.
I will make the the tutorial later. By the way, now we have the bookmark here.
So now you can just go on the coding canvas and you just click on this bookmark and buy like like magic this uh little window will appear. So now we have to go on pyicks code.py.com.
You can download the code for this demo on my GitHub repository as well and you install you have to import two two files csai_bridge.pai AI which is the code that intercepts the stream of data from the bridge running on your browser and then you can write any program for it. You just import this line from CSAI_Bridge import AI features and feature and then you you use the this object to get the detected class uh all the classes the confidence score or any coordinate from any like AI.get get cords and then the features you want like nose or let me show you nose or any any of them left here elbow eye hip knee shoulder whatever okay we we don't need this now so let me revert this and comment it out so now you connect to your py powered hub whatever you have I think all of them are compatible but I tested it on spike essential spike prime hub then you run the Okay, it's ready. Let me turn the program off now and we can disconnect.
Now we click on connect. The program has to be stopped for this to be uh found and uh paired to. We pair to it. Now the stream is already running. So if I do this, ah let me Okay, so now I can turn on the other camera. So now you can see it. So arms up, arms, uh up, right up, left up, both down. So you can see it works. So now that we install it and you can pause the stream at any point. So you see it's not working anymore. So you can I can start stream again and from here I can stop the program al together or start the program program again and start the stream again. You see?
Okay. And you can also disconnect. There is no way. Uh I mean this can be left here. And if you want to close this and not use it anymore, you just have to reload the page. Let's try it again with a full screen so you can see it better.
Up, down, right up, left up, and so on.
So it works. You can do it with any other pose classification. So you can also train other other classes. So you go here, arms down, you add another class, you add any classes you want. So you can train other classes and in the in the in the program you will just see more classes like class you get detected class you can have zero one two and you just have to match the class you train here uh on py start with zero here it starts with one you can rename them but the names won't be passed through the stream that's an advanced feature that you I might uh explain in another video but for now it's just about the setup and the initial demonstration. So now I would like to show you also how to use this on Android. So let me switch to Android. On Android you have to go to code.leucation.com and then you create a new project. You click on post classification. You wait for it to load.
Then you enable the webcam. Sorry, the camera of your phone. Hi. Here I am. So now um we have to reduce this one. And now how to enable the bridge. You cannot just go to here to the f um bookmark of favorites and click on the previously created bookmark. So I I suggest you create your bookmark on your computer first. So the best way is you go to your address bar and you type CSI not this page but the one you find on your favorite. So on if you type if you start typing you find your favorite here on the favorite bar. You click on that and you can see that the same windows as before will appear. So now I will can you you can see I I will stop the program on the hub. Click on connect.
You pair with it.
And now it's ready. So now you can see that it's working on your Android phone as well.
I hope you found this tutorial useful.
So you find all the links you need to do this under this video. Please like this video, share it, and subscribe. and see you on the next
Vidéos Similaires
Beyond Robotics | European Rover Challenge 2026
beyondrobotics
189 views•2026-06-01
Beatbot Sora70: JetPulse Technology and AI obstacle avoidance and navigation!
DroidModderX
26K views•2026-06-02
NVIDIA Just Unveiled a 75 DOF Humanoid Robot
DPCcars
74K views•2026-06-01
Tesla FSD 14.3.3 Hits Phoenix Streets - FIRST LOOK
anthonystesla
114 views•2026-05-29
NVIDIA's NEW Humanoid Robot Isaac Groot Revealed!
DPCcars
3K views•2026-06-01
Reachy Mini: the $300 open source robot you can actually hack — Andres Marafioti, Hugging Face
aiDotEngineer
662 views•2026-05-29
GTU ROVER | ERC 2026 Final Video
GTURover
401 views•2026-05-31
Tesla's $25 Trillion Dollar Problem | Will Optimus Win?
RyanShawtech
253 views•2026-05-30











