Pose Estimation on the Raspberry Pi 4!

The goal of this project was to how well pose estimation could perform on the Raspberry Pi. Google provides code to run pose estimation on Android and IOS devices - but I wanted to write python code to interface with and test the model on the Pi.

Here's a demo if you want to try things in your browser and with JavaScript!

https://storage.googleapis.com/tfjs-models/demos/p...

Additional resources:

• Medium Post:

  • https://ecd1012.medium.com/pose-estimation-on-the-...

• YouTube Video:

• Raspberry Pi 4 GB
• Raspberry Pi 5MP Camera (rev 1.3)
• LED
• 470 Ohm Resistor
• Small breadboard
• GPIO push button
• 3.5 Amp USB-C Power Supply

Command: git clone https://github.com/ecd1012/rpi_pose_estimation.gi...

Command: cd rpi_pose_estimation

Command: sudo apt-get update && sudo apt-get upgrade

Command: sudo pip3 install virtualenv

Command: python3.7 -m venv TFLite-venv

Command: source TFLite-venv/bin/activate

Download the following: https://github.com/ecd1012/rpi_road_object_detection/blob/main/get_py_requirements.sh

Command: bash get_py_requirements.sh

Command: sudo raspi-config

Go to interface Options and make sure the Pi Camera is enabled.

Command: wget https://storage.googleapis.com/download.tensorflow.org/models/tflite/posenet_mobilenet_v1_100_257x257_multi_kpt_stripped.tflite

Connecting the Camera for This Project: https://youtu.be/Zfmo3bMycUg

This will be used as input to kick off the program.

Connecting Push Button for this project: https://youtu.be/Zfmo3bMycUg?t=437

This LED will turn on to indicate when the program is running. Make sure you use a current limiting resistor with the LED!

LED Connection for This Project: https://youtu.be/Zfmo3bMycUg?t=353

After all your hardware and software is configured correctly run the following command:

Command: python3 TFLite_pose.py --modeldir notebooks/posenet_mobilenet_v1_100_257x257_multi_kpt_stripped.tflite --output_path pose_images

Where the --output_path you specify is where you want images saved.

The script will start running and wait for you to press the GPIO input button to start processing the video feed from the camera. Once you press the button, the green LED will turn on and the pi will start feeding and processing the video stream through the neural network. Processed images will be saved to the '--output_path' you specified over the command line, with a timestamped folder for each button press.