Project for PP RAI 2024

Core

• Event loop
• Wifi driver
• Cam driver
• Processing JPEG by edge detection algorithm
• Sending JPEG to Host
• Receiving JPEG from ESP32 via server.py

TODO

• Fix error related to Torch ai model arch: output_shape->data[shape_size] != dim_shape(1 != 0) Node STRIDED_SLICE (number 48f) failed to prepare with status 1

Installation Steps for Edge Detection AI Model on ESP

Follow the steps below to set up and run the edge detection AI model on ESP using ESP-IDF, TensorFlow, PIL, and Socket.

1. Install ESP-IDF

Follow the official ESP-IDF installation guide to set up the ESP-IDF development environment:

git clone --recursive https://github.com/espressif/esp-idf.git
cd esp-idf
./install.sh

2. Install Python Packages

Install the required Python packages, including TensorFlow, PIL (Pillow), and Socket:

pip install tensorflow pillow socket

3. Enable SPIRAM, AI model, setup wifi credentials

Run the following command to configure the project using menuconfig:

idf.py menuconfig

Ensure WPA2 PSK for WiFi Connection

4. Build, Flash, and Monitor

Build the project, flash it to the ESP, and monitor the output:

idf.py build flash monitor

5. Run Server

For capturing processed jpeg:

python3 server.py tensor

For capturing raw jpeg:

python3 server.py jpeg