I am trying to control a maze's ball using the vision information. I need to move the ball to its home position using the vision system's information. I am able to detect the edges of the route the ball has to adopt.
The algorithm has detected all the edges in the image. How may I extract the edges of the line the ball has to follow?
I did it using the video stream of Canny Edge detection example
I am not sure if I understand what you mean with "the edge of the line the ball has to follow". Is it a straight line? For detecting straight lines, you can try the function ImageProcessor::HoughTransformLines (there are two implementations, see the documentation). There is also an example IVT/examples/HoughTransformLines (resp. IVT/win32/HoughTransformLines). Of course there are also other approaches to line detection, which might be suitable for your application. The IVT, however, offers for line detection an implementation of the Hough transform only.
I am sorry for not being clear in my post. I am using IVT to solve a maze like the one shown below:
I am using stepper motors to operate the maze using the information from the vision system. I am trying thresholding and edge detection techniques. However I am confused what would be the best approach.
Your examples are really helpful.
I hope you understand that cannot help you solving a specific computer vision problem. However, I would not try to recognize the layout of the maze online, as it is fixed. This information should be fed to your system offline, so that you only have to do a localization of the boundaries of the maze and the ball online.
I apologize for not being specific. As of now, I am solving the problem with OpenCV as I am a novice with image processing. I am planning to port the libraries to IVT as I read that it is faster than the openCV.
IVT is not faster than OpenCV in general - it depends on the functions/methods you use. Furthermore, if speed is your issue, then the Keyetech Performance Primitives (see the topic "Highspeed" in the menu of the IVT homepage) might be of interest to you.
Good luck with your project!
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.