In this project I try to measure the pH value with litmus paper, a color sensor and machine learning. The color sensor is connected to a ESP8266 and sends a color measurement to a preconfigured Azure Machine Learning web service.
I started of with collecting the color data of the comparison strips and two measurements with the litmus paper in a CSV file. This file is used to train the Azure Machine Learning algorithm.
After training and comparing different algorithms I did a final run and deployed it as a web service.
A light/color sensor that reads the papers' color and loads that value to the web, so that the trained algorithm can deduce what the PH is from its training?
I'm 100% sure this works but isn't this just as good as using the exact same hardware and a lookup table?
Its pretty clever and a bit over kill i think, but i like the idea of a PH web service... current PH sensors are a pain to use and im sure you are on the right track to something.
It could be done with a lookup table but the colors of the comparison strips are a bit different from the actual colors of the litmus paper. You could compensate for this but sometimes it's just fun to do it differently :) but it's indeed overkill.
A light/color sensor that reads the papers' color and loads that value to the web, so that the trained algorithm can deduce what the PH is from its training?
I'm 100% sure this works but isn't this just as good as using the exact same hardware and a lookup table?
Its pretty clever and a bit over kill i think, but i like the idea of a PH web service... current PH sensors are a pain to use and im sure you are on the right track to something.