I want to detect fingers tips movement with UNO Q and camera
can I install CV2 and MediaPipe in the UNO Q
Arduino board has a microcontroller. You cannot install any library in any UNO board. You have to use firmata library for your purpose.
https://docs.arduino.cc/retired/hacking/software/FirmataLibrary/
You have to download and install firmata library. Then you need to go to
Examples-->Firmata-->standardfirmata. You have to upload standardfirmata in your UNO board. Then you need to write a separate python script. In that python script you need to include all necessary libraries including CV2, mediapipe, pyfirmata and others. The script will integrate the arduino board to python.
Hi @loveforcircuits. This discussion is about the new Arduino UNO Q. In addition to its STMicroelectronics STM32U585 microcontroller, this board has a microprocessor running Debian Linux. That Linux machine acts as a standard computer, so can indeed run Python applications based on OpenCV.
Arduino provides a more powerful communication interface between the Linux machine and the microcontroller:
https://docs.arduino.cc/tutorials/uno-q/user-manual/#bridge---remote-procedure-call-rpc-library
@ptillisch : I see. So, is it possible to install CV2 and MediaPipe in the UNO Q directly?
as far as I know this will work with the UNO board that has microcontroller only but UNO Q has also microprocessor has Debian Linux. and has Python 3.13.5
It is definitely possible, as others have already reported doing so:
When we are referring to installation on the primary Linux machine, it is ideally no different from installation on any standard Linux PC. The only caveat is that you may find that some software has components that are specific to x86 CPU architecture, and thus not compatible with the ARM CPU of the UNO Q's Dragonwing QRB2210 microprocessor. However, ARM CPUs are fairly commonplace in this day (partly thanks to the long term popularity of Raspberry Pi, and also due to even consumer PCs increasingly adopting ARM processors). So in many cases you won't encounter any difficulty in regards to processor architecture compatibility.
The task becomes a bit more complex if we are talking about using the software in an Arduino App. The reason is that the Python script component of the App, which runs on the microprocessor, is running inside the isolated environment of a Docker container. So it is limited in how it can interact with software installed on the primary Linux machine.
In general, the intended approach to this type of project is to provide the additional resources to the App through a secondary container created by adding a "custom Brick" to your App:
https://docs.arduino.cc/software/app-lab/bricks/custom-bricks/
I would start by checking to see if an existing image is available that provides the required capabilities. Docker Hub is a good place to discover such images:
Failing that, you can create your own custom container or image. There is some information about that in the custom Brick documentation, and the "Bricks Architecture and Configuration Reference":
https://docs.arduino.cc/software/app-lab/bricks/bricks-reference/
You might also be interested in the discussions on the subject of custom Brick and Docker images that occurred in this forum topic: