Seeed Studio Grove Vision AI firmware

Edge Impulse enables developers to create the next generation of intelligent device solutions with embedded Machine Learning. This repository contains the Edge Impulse firmware for the Seeed Studio Grove Vision AI module. This device supports all Edge Impulse features, including ingestion, remote management, and inferencing.

See our tutorial on how to use the module with Edge Impulse Studio

NOTE

If you are using the Grove Vision AI module with Edge Impulse for the first time, upgrade the firmware on the USB-UART BL702 converter.

Download the following firmware from Edge Impulse CDN server.
Follow the instruction to flash the BL702 firmware.

Source code organization

You can find the main firmware source code here:

Synopsys_PA8535_EM9D_DFSS_SDK_3.3/app/scenario_app/edge_impulse_firmware
To build the firmware, use the following makefile:

Synopsys_PA8535_EM9D_DFSS_SDK_3.3/makefile
If you want to deploy your model (exported as a C++ Library from Studio), extract the exported zip file content into:

Synopsys_PA8535_EM9D_DFSS_SDK_3.3/app/scenario_app/edge_impulse_firmware/ei-model

Reporting issues

This repository is a snapshot of the Edge Impulse internal code base, therefore PRs won't be merged. If you find any bugs or want to contribute use our Developer Forum to submit them.

How to build the firmware?

You can build firmware using the provided Dockerfile (the most straightforward solution). Another option is a native local build.

Build firmware using Docker

Make sure you have Docker installed.
Build the container:
```
docker build -t seeed-vision-ai-build .
```

Run the build script using

docker run --rm --interactive -v $PWD:/app seeed-vision-ai-build /bin/bash build-firmware.sh

You can find the resulting firmware in firmware.uf2 file.

Build using native tools

We tested this build method on Ubuntu 20.04 and Fedora 36.

Install the required tools

sudo apt install make
sudo apt install python3-numpy

Download and install the GNU toolchain for ARC processors

cd ~
wget https://github.com/foss-for-synopsys-dwc-arc-processors/toolchain/releases/download/arc-2020.09-release/arc_gnu_2020.09_prebuilt_elf32_le_linux_install.tar.gz
tar -xvf arc_gnu_2020.09_prebuilt_elf32_le_linux_install.tar.gz
export PATH="$HOME/arc_gnu_2020.09_prebuilt_elf32_le_linux_install/bin:$PATH"

Build the firmware and convert it into UF2 format

cd firmware-seeed-vision-ai-internal/Synopsys_PA8535_EM9D_DFSS_SDK_3.3
make
make flash
python3 tools/ufconv/uf2conv.py -t 0 -c tools/image_gen_cstm/output/output.img -o firmware.uf2

Flashing the board

Below is the procedure to flash the firmware on the Grove - Vision AI Module.

Connect Grove - Vision AI Module to the host PC via USB Type-C cable
Double-click the boot button on Grove - Vision AI Module to enter mass storage mode
After this, you will see a new storage drive shown on your file explorer as GROVEAI
Drag and drop the previous firmware.uf2 file to GROVEAI drive

Once the copying is finished, GROVEAI drive will disappear.

How to view the camera stream?

Install Edge Impulse CLI tools
Connect your board to the PC, open the terminal/command line, and run:
```
edge-impulse-run-impulse --debug
```

You should see an output similar to the one below. Open the link

Edge Impulse impulse runner v1.15.1
[SER] Connecting to /dev/tty.usbmodem21401
[SER] Serial is connected, trying to read config...
[SER] Retrieved configuration
[SER] Device is running AT command version 1.7.0

Want to see a feed of the camera and live classification in your browser? Go to http://127.0.0.1:4915

Open the displayed link in a browser, and you will see the preview:

n-miyamoto/firmware-seeed-grove-vision-ai