Visual Servoing In Torch Transferring Of Robot

This article comes from a sub-task of 863 program——the research of compliance control techniques on apery robot. We study the uncalibrated visual servoing control by eye-in-hand robotic system to realize static visual tracking of the torch on PA10-6CE robot.The assignment of this article is that ,during the process of handing over and taking over the torch by the robot ,first we use the camera and image collection card to transfer the visual information that gets from the camera into digital visual information,and then recognize the torch from the robot's working environment. We choose the color features and geometry shape features of the torch as dependence for the torch's recognition,and compose the two kinds of features into significative image features.We adopt a series of image processing arithmetics to recognize the torch , then distill useful image features: the barycenter of the object. We adopt the uncalibrated visual servoing method to fulfill the online recursive estimation of the pseudo-inverse matrix of image Jacobian , and computer the change of the robot's joint angles , then feed back to the PA10-6CE robot controller to make the robot's joint angles move, and at last lead the robot to track the torch in the image plane by adjusting the value of the pseudo-inverse matrix of image Jacobinan so that to accomplish the static visual servoing of the industrial robot.