Visual servoing based self-calibration of robotic inspection system using rigid body transformation parameters

Abstract

Calibration plays a vital role in maintaining the positing accuracy of the deployable type of industrial robots. A fully automatic deployable type of robotic manipulator has been developed for inspecting the steam tubes of a steam generator (SG). It uses damped least squares-based kinematic algorithm to position the tube inspection probe exactly over the centre of individual steam-tubes and pushes probe into the tube for detecting any flaws. As direct global referencing is not possible inside the space constrained SG, a novel robot calibration methodology has been developed that uses vision data for recognising the reference V-patterns of the steam tubes arrangement through a vector approach. Robotic arm kinematic errors are corrected using rigid body parameters estimation by exploiting the joint motor encoders used for sensing the pose. Further, using visual servoing technique, high order positioning accuracy of inspection probe has been achieved to reach every steam tube centre for inspection.