I am reading tutorials on visual servoing using Hand book of robotics and there it is mentioned that in order to control camera's six dof we need atleast 3 points. I didnt quite get the idea behind it? can anyone please explain. Thanks
We need to be able to track 3 points, (sometimes called features), on a solid body to estimate its 6dof pose relative to the camera.
Each has an estimated position in the 2D camera image.
If you imagine having only two points visible on the object, you can rotate the object around the line (axis) joining the two points and the 2D position in the image will not change.
Since there is no change in the position of the points in the image for this rotation, we need one more point, that is not on the line, to estimate the pose.
You can also just count the Degrees of Freedom if you understand what is independent in your problem domain. In this case, each point has 2DoF, so you need 3 of them to get 6DoF.
Note that in practice we use more than 3 points to both average errors and to handle the fact that it is not always easy to identify which point is which.