Real-Time Object Detection on Raspberry Pi 5 with Coral Edge TPU

This project enables real-time object detection on the Coral Edge TPU integrated with a Raspberry Pi 5. With the FastAPI server, you can easily send images and receive a list of detected objects accompanied by bounding boxes, confidence scores, and labels.

Key Features

• Fast and efficient object detection: Leverages the Coral Edge TPU for hardware-accelerated machine learning.
• Easy-to-use API: Simple image submission and result retrieval.
• Example usage and visualization: Provides clear examples for using and understanding the results.

Prerequisites

• Raspberry Pi 5 with Coral Edge TPU M.2 attached to PCIe slot.
• Compatible Operating System:
  • Debian GNU/Linux 12 (bookworm) 64 bit
• Linux kernel 6.6.20+rpt-rpi-v8. Check with uname -r.
• Python 3.9.16
• Linux PCIe Driver for Coral Edge TPU M.2 documentation

Requirements

To install Docker, follow the instructions in the official Docker documentation.

sudo apt install devscripts debhelper dkms -y

echo "deb https://packages.cloud.google.com/apt coral-edgetpu-stable main" | sudo tee /etc/apt/sources.list.d/coral-edgetpu.list

curl https://packages.cloud.google.com/apt/doc/apt-key.gpg | sudo apt-key add -

sudo apt-get update

Install libedgetpu

sudo apt-get install libedgetpu1-std

Running the Server

Pull Docker Image and start the FastAPI server:

docker pull ghcr.io/ajmalrasi/object_detection_tpu:main

docker run --device=/dev/apex_0:/dev/apex_0 -v /usr/lib/aarch64-linux-gnu:/usr/lib/aarch64-linux-gnu:ro -p 8000:8000 -it <image>

Running the Client

Install Requiremnets

pip install -r requirements.txt

Edit the Camera Index and Run Script

python3 client.py