Ask Your Question

streets segmentation from aerial view via opencv

asked 2013-03-13 05:02:58 -0500

matteo gravatar image

hi, i have a video of a real city-environment made by an aerial camera whose frames i'd like to extract only the streets shapes from. does an algorithm (or a well known series of opencv functions) exist for my purpose? thanks!

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2013-03-13 05:20:49 -0500

This is a question that is easy to answer. NO there is no built-in algorithm specifically designed for detecting roads. HOWEVER, there is OpenCV functionality that could get you to the point you want to.

For detecting roads there are different approaches possible, first you should read papers concerning this topic, for getting an idea of the algorithms around.

Basically these steps could be followed - Perform edge detections on image - Try to apply a linear Hough transform in order to find lines in the image - Then define how far lines should be apart (a fact that is known, since aerial images are taken from a constant height). - Use the knowledge to remove unwanted lines that are not close enough and could not be grouped.

This is just an idea sprung from my mind. Saw some presentations about this at the last VISAPP conference in Barcelona. This is still a hot research topic in computer vision.

OpenCV provides functionality for edge detection and hough transforms. Go through the docs for that.


edit flag offensive delete link more
Login/Signup to Answer

Question Tools

1 follower


Asked: 2013-03-13 05:02:58 -0500

Seen: 541 times

Last updated: Mar 13 '13