Looking for some example scripts

[mentalrage]

So I've been programming simple things for awhile (wow mods and such) and I wanted to do a program for another game that can read information from the game window. This game shows the players cords and the current map name on the screen and what I wanted to do first of all was to be able to read what that information using pixel detection. I want to take screen shots of the various individual numbers that show up and the map names that appear. I then want to make a script that compares the individual bmp to a certain rectanglar area on the window. Basicly take the bmps of 0 - 9 then compare it to the number thats showing on screen to find a 100% match. I just don't know where to start. So I was wondering if there was any scripts that do this already I can look at or possibly some links to sites explaining it. Thanks.

[Angoth]

So I've been programming simple things for awhile (wow mods and such) and I wanted to do a program for another game that can read information from the game window. This game shows the players cords and the current map name on the screen and what I wanted to do first of all was to be able to read what that information using pixel detection. I want to take screen shots of the various individual numbers that show up and the map names that appear. I then want to make a script that compares the individual bmp to a certain rectanglar area on the window. Basicly take the bmps of 0 - 9 then compare it to the number thats showing on screen to find a 100% match. I just don't know where to start. So I was wondering if there was any scripts that do this already I can look at or possibly some links to sites explaining it. Thanks.

IMHO, you'll have a really hard time doing things this way. It's not the 'foreground' match that's the problem, it's the underlying stuff on the screen that happens to be there that will preclude a 100% match to a stored BMP. If you could positively say everything but the number was black, then, it could work. I just think that pixel detection for this job is inefficient, at best.

Try OCR for that job.

Have a look at this thread:

Tesseract OCR UDF

Let me know your testing results.