In this video, I show you how object detection, classification, and tracking work on a Raspberry Pi 5 using YOLO and the Hailo 26 TOPS AI HAT. You'll get to see these AI tools in action, detecting and tracking objects accurately and efficiently.
I also walk through how I've structured my application using modular GStreamer pipelines. This design lets multiple processes easily access the same camera feeds, keeping things organized and scalable.
Plus, I share some custom Python code I've created that integrates directly into the GStreamer pipeline. This code identifies specific tracked objects and then sends relevant data to AWS IoT Core via MQTT.
If you're curious about object detection, Raspberry Pi projects, or how to connect your devices to the cloud, you'll find practical guidance here.
What you'll learn:
- How YOLO handles object detection and classification
- Setting up real-time object tracking on Raspberry Pi 5 with the AI HAT
- Creating modular GStreamer pipelines
- Integrating Python code with GStreamer
- Sending data from Raspberry Pi to AWS IoT Core via MQTT
00:00 Intro
00:32 Demo
03:52 How the application is organised
08:17 Producer camera
11:36 Consumer Tracking
23:33 Consumer fpsdisplaysink
26:22 FAQs
All the code shown:
https://gist.github.com/jameselsey/25016415575675b89e11d6784e52dff6COCO common object list:
https://github.com/ultralytics/ultralytics/blob/main/ultralytics/cfg/datasets/coco.yamlWant to say thanks? I have a book list with lots of things I want to learn:
https://www.amazon.com.au/hz/wishlist/ls/1WT2E6DONHH8Z?ref_=wl_share#RaspberryPi5 #YOLO #ObjectDetection #RaspberryPiProjects #AWSIoT #MQTT #GStreamer #Tutorial