MATLAB Answers

Computer Vision - Feature Detection and Tracking - Tracking a RC car in a video

2 views (last 30 days)
Mark
Mark on 13 Jun 2012
I would like to be able to track three different features of a RC car in an avi video. The camera is fixed. I hope to be able to calculate the actual position of the car once I am able to track the car in the video.
The main problem is with not knowing much about feature detection and tracking. I have the student version of matlab but don't have the computer vision toolbox. Is there a way I can do this feature detection and tracking without using the computer vision toolbox?

Answers (1)

Geoff
Geoff on 13 Jun 2012
Of course, but you might have to implement a bunch of complex algorithms yourself. If you're not familiar with vision processing, you might struggle.
What features on the car are you talking about? Do they have a unique shape and/or colour? Are they always visible? As for recovering the position of your car, your success will depend on how well you are able to calibrate your camera and/or how flat the ground is, how high your camera is, lighting conditions etc....
Simplest way to track a moving object in a stationary camera is by frame differencing. But the complexity of your solution depends on what else is moving in the scene. I would start with blob tracking, and if that doesn't suit, maybe optical flow.
It might help if you post up some frames from your video.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by