Classify Objects on Pi with Edge Impulse
Вставка
- Опубліковано 6 сер 2024
- Embedded image-based machine learning is a technology paradigm that is becoming more and more useful, especially in the field of IoT. As deep learning networks get simpler and smaller, we can expand the power of ML to devices like the Arduino or the Raspberry Pi, supercharging their utility in computer vision tasks in particular.
In this workshop, we’ll show you how to train your Raspberry Pi to detect and classify objects using the very impressive and easy-to-use Edge Impulse web platform for TinyML (via the BalenaCloudOS, which has pre-built API calls to Edge Impulse).
Prerequisites:
- Raspberry Pi
- Smartphone
- Laptop
- Pi Camera or a USB webcam
Link to deploy the application: dashboard.balena-cloud.com/de...
-----------------------------------------
To learn more about The Assembly’s workshops, visit our website, social media or email us at workshops@theassembly.ae
Our website: theassembly.ae
Social media:
-Instagram: / makesmartthings
-Facebook: makesmartthings
-Twitter: / makesmartthings
-----------------------------------------
0:00 1. Intro
1:59 2. Workshop overview
3:52 3. Go to the Edge Impulse website
5:03 4. Capture photos for the model
12:44 5. Generate features of the images
14:57 6. Train & test the model
20:30 7. Export the model
22:13 8. Create a new application on Balena
24:45 9. Add a new device (Pi) & download balenaOS
29:15 10. Test the output
#EdgeImpulse #RaspberryPi #BalenaCloud - Наука та технологія
Good video! Thank you for sharing! :)
How can I use the Edge Impulse model in a Python application running in the RPi4, such that I can further send the counted detected objects (along with its percentage accuracy) to another device like an Arduino or a Windows 10 PC?
Sir, when we copy the local ip address and paste it in browser it is showing that "this site cant be reached"
Hi. I am trying to use a v1.3 camera and I have the error: Failed to initialize linux tool Error: Cannot find any webcams, run this command with --disable-camera to skip selection
at /usr/lib/node_modules/edge-impulse-linux/build/cli/linux/linux.js:423:23