I’d say that while not technically impossible, it’s not a good fit for a Particle Photon/Electron/Core.
The camera options are limited, and the more substantial limitation is the amount of free RAM to process the image.
I’d go with a Raspberry Pi instead. There’s a nice, inexpensive, camera module, and there are good open source image processing libraries for Linux that will help in reading the numbers from the dials.
I agree with @rickkas7, a camera is not really suited for a photon or similar. Perhaps a Pi might help but still, image recognition is no easy task. Perhaps identifying each 0.1m3 would be easier with a color sensor (might be too wide) or laser or something along that line. That would be perfectly doable with a particle board
Ofcourse, you’d still need to manually add the starting amount.
And while I realize you can’t modify your water meter, the way it’s typically done in industrial uses is to put a tiny magnet on the least significant dial, then place a hall-effect sensor on top of the clear cover to measure when the dial makes a complete revolution. That task is easily handled by a Photon.
I bet you could use a color sensor. That one has a white LED on the board as well. Just set it over the least significant dial and you might be able to see the red pointer as it passes under the sensor and just count the number of revolutions.