Tracing the grid...
I am out of time right now, because I have to do a research paper, since Eileen ripped me off.
But this is how far I got with my algorithm today:
Finding the 1st point of the left vertical line and tracing a view points down:
Automatic detection of the end of the line:
A second line that is not traced because its not the left one. Also notice the many dots on the bottom. When no further point is found, the steps of the "search radius" get smaller to reach the end more percisely:
Also very close points at the almost horizontal piece of the line, otherwise the program would have been able to follow the line to the right:
Here the trace is now centered, before, it was always at the very left of the line.
Since I am scanning from top to bottom, the first point could be detected on the wrong line if it is higher than the other one. This was a harder one to solve.
You can see one red dot on the very top of the right line, which was the first point detected; however the program found the left line as soon as it came in "scan range":
Horizontal lines didn't seem to be causing any problems. If you noticed the last point being too far down, that was a bug that I fixed later on.
...Hours go by...
All lines correctly traced:
Here I got those crazy lines going all over the place really fast, while trying to draw lines on the web cam.
The traced lines projected onto the web cam image:
Something got messed up here too, I asume it was the algorithm that detects the background (you see alot of white spots where there arent even any lines).
Here is a decend one (other than me goofing off):
I don't even know what happend here, but it was frustrating:
And this one didn't work at all:
I might have no other option than to recode the background algorithm, cause this sucks.
Thx for reading :)
But this is how far I got with my algorithm today:
Finding the 1st point of the left vertical line and tracing a view points down:
Automatic detection of the end of the line:
A second line that is not traced because its not the left one. Also notice the many dots on the bottom. When no further point is found, the steps of the "search radius" get smaller to reach the end more percisely:
Also very close points at the almost horizontal piece of the line, otherwise the program would have been able to follow the line to the right:
Here the trace is now centered, before, it was always at the very left of the line.
Since I am scanning from top to bottom, the first point could be detected on the wrong line if it is higher than the other one. This was a harder one to solve.
You can see one red dot on the very top of the right line, which was the first point detected; however the program found the left line as soon as it came in "scan range":
Horizontal lines didn't seem to be causing any problems. If you noticed the last point being too far down, that was a bug that I fixed later on.
...Hours go by...
All lines correctly traced:
Here I got those crazy lines going all over the place really fast, while trying to draw lines on the web cam.
The traced lines projected onto the web cam image:
Something got messed up here too, I asume it was the algorithm that detects the background (you see alot of white spots where there arent even any lines).
Here is a decend one (other than me goofing off):
I don't even know what happend here, but it was frustrating:
And this one didn't work at all:
I might have no other option than to recode the background algorithm, cause this sucks.
Thx for reading :)