Research On Robotic Visual Servoing System

In industrial processes, robots can replace humans do a lot of work if the robot supporting the corresponding sensor, such as: workpiece pickup and sorting, palletizing,welding tracking, etc, has a very large application value. Industrial robots in the implementation of production tasks, need to identify, locate and grab the target object, and binocular stereo vision can provide a wealth of information,so that industrial robots have the ability to perceive the surrounding environment. Research on robot visual servo system, based on the location, grab the target object, eye-in-hand calibration, target identification and location, six degrees of freedom manipulator motion control and related algorithms can be verified in detail, provide a theory verification platform for industrial robot visual servoing system grab workpiece and other objects, far-reaching significance. The main contents are as follows:First,calibration the relationship between the manipulator and camera. Introduced the method described the position and orientation in the space coordinates and coordinate transformation between two space coordinate system. The hand-eye equation has been established for the eye-in-hand system, summarized the procedure of the traditional methods for the hand-eye equation and analyzed the error source, and then introduced the optimization method for solving hand-eye equations. The hand-eye calibration algorithm and experiments has been designed for the robot visual servoing system in this paper and the transition matrix between the manipulator and camera has been acquired by experiments.Secondly, the target object in space feature extraction and three-dimensional position measurement algorithm has been studied. A target recognition algorithm based on contour,shape, color and other characteristics have been proposed for recognition of the target object.To measure the position and orientation of the target object, camera imaging model and binocular vision measurement principle has been studied. A software combined with target recognition algorithm and binocular vision measurement principle for real-time image processing and measuring the pos of the object has been written,the object recognition and positioning experiment has been carried out, the experimental results show that the target recognition algorithm and binocular vision measurement principle are effective for identification and location of the target object.Thirdly, the robot motion control algorithm has been launched. The model of the six degrees of freedom manipulator has been established based on the D-H method,according to the transformation matrix between adjacent two coordinate systems derived forward kinematic equations, and using the Robotic Toolbox verify the correctness and rationality of the model and the forward kinematic equations. Inverse kinematics algorithm of the robot based on the algebraic method has been derived, which has been achieved in the Microsoft Visual C ++ 6.0 programming environment, and the correctness of the algorithm has been proved by forward kinematics equation.Finally,Robot visual servoing system grab and setting target object experiments has been designed. The robot visual servoing system has been set up and the system software has been written in Microsoft Visual C++ 6.0 compiler environment, the target object placed on the table has been grabed and placed by the system, which has proved that eye-in-hand calibration, identification and location, robot kinematics and other related algorithms are reasonable and practical.