Eye-to-Hand

A camera-mounting scheme where the camera is fixed outside the robot and doesn't move with the arm, plus its matching calibration.

Eye-to-hand is one of the two hand-eye calibration configurations: the camera is fixed — on a tripod, table edge, or stand — and doesn't move with the arm, as opposed to ‘eye-in-hand,’ where the camera is mounted on the end-effector. Eye-to-hand requires the fixed transform from the camera to the robot's base (the camera extrinsics); once known, it lets objects seen by the camera be mapped into coordinates the arm can act on. Calibration works by fixing a calibration board or an ArUco marker to the end-effector, moving the arm through several poses while the camera photographs each one, and solving an equation of the form AX = XB, commonly with the Tsai-Lenz algorithm; MoveIt and easy_handeye both offer ready-made tools for this. Its advantage is a stable view that can see the whole worktable; its drawback is that it's easily blocked by the arm itself, and bumping the camera means recalibrating.

ExampleA fixed depth camera on a tripod at the table's edge looks down at the workspace. A ChArUco board is clamped to the arm's end-effector and moved through a dozen or more poses, each photographed once, to solve for the camera's pose relative to the base. Afterward, cup coordinates the camera detects can be sent directly to the arm to grasp.

Also called

Fixed Camera Calibration, Eye-on-Base, Eye-to-Hand Calibration

Related

Hand-Eye Calibration · Eye-in-Hand · Camera Extrinsics · Calibration Board · Third-Person Camera · ArUco Marker

Sources

easy_handeye: eye-in-hand and eye-on-base calibration (GitHub)
MoveIt 2 Hand-Eye Calibration Tutorial