Yesterday Duccio Fioravanti has succeeded in tracking a planar target using image based visual servoing with a kuka robot.
He used the CaptureCamera Component to get the images from a AVT Guppy firewire camera which was connected to the robots end effector. He created his own components for the control in the image and used the CartesianControllerVel and CartesianEffector to control the robot. He could track an object and made the robot follow it.
Also planning in the image
Duccio has added a planner to his image based vision control. Now he can also have stable control in the image when having a big difference between the desired and the initial view of the object.
The next step, he says, is adding online estimation of the depth ratio between the different tracked points.
Ruben