I have a project that needs to close a relay (unlock a door) based on a text string. FYI I would need to read the text string from about 2 meters away, maybe further, however for testing purposes, this isn’t important just yet.
I believe? that both Arduino and Raspberry Pi can do this. However, I am not sure which is suited best.
Also, I would like the software to run locally (no third-party interface).
Also, what is the difference between text recognition and numberplate recognition? it seems like they would basically do the same thing.
Therefore, what hardware and software combination would be recommended?
While some Arduino-produced boards do have a decent bit of power, most run on older 8-bit microcontrollers that have pretty much no chance of processing high-resolution images to recognise text at 2m out. A Pi is a much better option, plus it’s a lot easier to get images into (USB or CSI cameras)
Generally OpenCV2 is a tinkerer’s first port of call for image recognition. There are tons of guides on text recognition, and methods differ based on how “machine readable” the text is. More info here:
As for the difference between number plate and text recognition, I’d say a number plate detection system would be looking for certain shapes or combinations of numbers/letters in order to zero in on a number plate and ignore any other text. If you have more control of what the text looks like, it may be more accurate (i.e. ignoring logos on shirts, boxes, whatever).
Perhaps a QR code or barcode system would be more appropriate? Can you share any details about the motivation behind the project?