Click here to Skip to main content
15,886,258 members
Please Sign up or sign in to vote.
5.00/5 (3 votes)
See more:
I am working on an application that uses camera(s) to detect while lane lines on different surfaces adn track them, and eventually give directions to the vehicle (two wheeled robot) so that the vehicle stays in between the lines. So far I managed to write the program so it takes color video feed and converts it to black and white, and filters the edges so it detects the edges of the lanes. I don't know where to go from here. Any suggestions?
Posted

I agree with Uffe Rask, this sounds like an interesting project. Now, I know very little about Robots, modern ones anyway, but I googled for line following robot algorithm and got lots of hits that sounded to me like they might be of help. Perhaps if you do the same (always assuming that you haven't already done so), you might find something that you can adapt.

 
Share this answer
 
You only need one camera to achieve what you want. There was an old (~ 10 - yrs) DARPA(?) project which used an artificial neural network to control the steering as you described. My MSc dissertation (a game playing virtual "tank") was influenced by it.

The basics were as you described, take the video feed, then whizz up the contrast to black and white. You then end up with the white lines as two edges of a "trapezoid". The neural network was trained so that the car was kept in the centre of the trapezoid (this is the hard part).

There is a link here http://society.kisti.re.kr/~Eksae/_notes/data/pdf/v3n2_5.pdf[^]

to a similar project.

This will be an interesting project!
 
Share this answer
 


CodeProject, 20 Bay Street, 11th Floor Toronto, Ontario, Canada M5J 2N8 +1 (416) 849-8900