• DepthAI
  • Where can I found the source code of the firmware / depthai-device-side?

Hi jpizarrom ,

So we cannot release the firmware as it is based on tools that Intel considers proprietary. If this ever changes we will be able to open-source the firmware.

We were able to write the host side and the ESP32 from scratch not using any Intel tools, and that is why that can be open-sourced. And we made a module for the Myriad X so that we could open source the rest of the hardware.

Overall

  1. https://github.com/luxonis/depthai-hardware - DepthAI hardware designs themselves.
  2. https://github.com/luxonis/depthai - Python Interface and Examples
  3. https://github.com/luxonis/depthai-api - C++ Core and C++ API
  4. https://github.com/luxonis/depthai-ml-training - Online AI/ML training leveraging Google Colab (so it’s free)
  5. https://github.com/luxonis/depthai-experiments - Experiments showing how to use DepthAI.

Embedded Use Case

  1. https://github.com/luxonis/depthai-spi-library - SPI interface library for Embedded (microcontroller) DepthAI application
  2. https://github.com/luxonis/depthai-bootloader-shared - Bootloader source code which allows programming NOR flash of DepthAI to boot autonomously
  3. https://github.com/luxonis/esp32-spi-message-demo - ESP32 Example applications for Embedded/ESP32 DepthAI use (e.g. with BW1092)

What are you hoping to do with the firmware?

Perhaps we already can make happen what you are looking to accomplish. For example we will be supporting microPython soon (see here)

And we also have a version with an integrated ESP32, which allows you to then have your own custom logic/etc. on the ESP32 and actuator drive, sensor integration, etc.:
https://github.com/luxonis/depthai-hardware/tree/master/BW1092_ESP32_Embedded_WIFI_BT

And of course all the source code for the ESP32 we can open source (and have done so) because the ESP32 toolchain is fully open source. See here for some details and how to integrate microROS for example here

Thoughts?

Thanks,
Brandon

    Brandon ,
    Thanks a lot,

    I hope that I can implement post-processing on the device side, such as decoding. I saw that decoding for ssd and yolo is also available on device side, but I was hoping to be able to implement my own decoding as well, so that more models can be supported.

    Thanks

      jpizarrom

      Got it, thanks! Yes, you will be able to do this in microPython in our Gen2 pipeline builder. We expect to have this possible late December this year.

      This will allow you to write microPython code to do the decoding on-device as you describe.

      Thanks again,
      Brandon