Point Correspondence in Image-Based Visual Servoing


It's going to be useful to say a little bit more about the point correspondence problem.

Here we have again the initial view of our triangular object and here we have the final view, or the desired view, of our triangle object. 

When the points move in the image plane it's possible for them to move like this and this is the example that we just looked at a moment ago. But it's also possible for the points to move like this and if they did that then the motion of the object in three dimensional space is going to be much more complex. It's almost going to require the object to flip around in the three-dimensional world.

Another way we could define the correspondence is like this and again it involved a different kind of three-dimensional motion.

And finally we could have this particular way of point moving on the image plane.

Each different way of associating a yellow dot with a red dot is going to result in a different motion in the three-dimensional world. The net result will be the same in the image plane but the actual three-dimensional motion will be different.

A critical part of a visual servoing system is establishing correspondence between points in the scene observed by the camera, and points in our desired image of the scene.

Professor Peter Corke

Professor of Robotic Vision at QUT and Director of the Australian Centre for Robotic Vision (ACRV). Peter is also a Fellow of the IEEE, a senior Fellow of the Higher Education Academy, and on the editorial board of several robotics research journals.

Skill level

This content assumes high school level mathematics and requires an understanding of undergraduate-level mathematics; for example, linear algebra - matrices, vectors, complex numbers, vector calculus and MATLAB programming.

More information...

Rate this lesson


Leave a comment