Ask Your Question

Best way to detect wheels passing on video

asked 2017-08-28 16:13:52 -0500

updated 2017-08-29 05:13:57 -0500

berak gravatar image


I'm new to OpenCV but I've managed to get a few basic functions working like transparent overlay and video output etc.

My small project is to detect the speed of cars going past my house by processing MP4 videos from the same day. I figured the best step is to create a detection area as I have on the image and detect the wheels as they pass through.

So my question is, what is the next best thing for me to investigate so I can detect wheels in the video stream?

I'm using OpenCV3 and Python (Spyder app within Anaconda2).

See attached image for a still of the video feed.


edit retag flag offensive close merge delete



take a look at bg_subtraction

sturkmen gravatar imagesturkmen ( 2017-08-28 16:26:42 -0500 )edit

For the wheel detection you can try a Hough Circle Transformation

VxW gravatar imageVxW ( 2017-08-31 04:24:07 -0500 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2017-09-02 09:04:57 -0500

updated 2017-09-02 09:20:44 -0500

Let's try with a cascade classifier (HOG or LBP), it should be easy to build the dataset and test the cascade.

Take a look here to get the best performance: vision-ary project

edit flag offensive delete link more
Login/Signup to Answer

Question Tools



Asked: 2017-08-28 16:13:52 -0500

Seen: 333 times

Last updated: Sep 02 '17