Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Orientation of a known object in an image

I have a picture with a known, simple object, e.g. chessboard. The camera can have different angles, distances etc.

What is the "standard" object detection algorithm to detect the object's position, angle and tilt? From those three, position is probably the simplest one, later rotation and the tilt..

I am aware that it is not an easy task, and there might be plenty of algorithms that bite this problem, yet for an object known beforehand there might be some commonly used solution (like eigenfaces for face recognition)

like image 556
Jakub M. Avatar asked Oct 10 '22 09:10

Jakub M.


1 Answers

OpenCV has an implementation for this called findChessboardCorners. Here is the code they use to detect it. They also have a whole slew of functions for camera calibration and determining the extrinsic parameters. Here is one of their samples for calibrating a camera.

Finally, Google Scholar has many papers available on this topic.

like image 157
mevatron Avatar answered Oct 13 '22 11:10

mevatron