Visual control architecture of servoing systems based on image moments

In this paper an image moments-based visual control architecture of servoing systems is presented. As visual features, image moments are analyzed to establish performances in different types of servoing applications. An image-based visual control scheme for 6 d.o.f. manipulators is presented. Robot dynamics are modeled as a Virtual Cartesian Motion Device. The designed control law ensures convergence towards zero for the image plane error, convergence obtained from driving each decoupled joint of the manipulator robot. A Simulink implementation was proposed, that represents an extension for an existent Matlab servoing toolbox. Performances for point features-based and image moments-based control laws were analyzed.