Corrects the transformation of the robot's camera with respect to the robot's arm based on input from the camera, by looking at edges in the image as compared to the edges in a rendered image of the arm at its current joint angles.