A brief update on this program. Bruce has been doing a fantastic job testing the program on his Maggiolino roaster. This has resulted in a significant improvement in the reliability of the program. But we still have issues to resolve for it to work reliably (without occasional false readings) on his roaster. Hopefully we will have success soon.
The above screenshots are of the program working on my Hottop P which has an LCD display. I have now unpacked and cleaned my new Hottop B which has a very different LED display. I thought I would give the program a try on that as I ran the roaster up to temperature empty. It worked perfectly see the following screenshots:
So we can now say the program can read Hottop B and P data without problems and also reads my digital multi-meter without problems.
The Hottop B temperature display captured by the webcam.
The program capturing and processing the image.
The new seven segment optical character recognition engine I have written to improve reliability of readings.
The chart produced heating the roaster without beans.