Research On Pose Measurement Method Of Target Based On Robot Hand-eye Vision

The progress of science and technology leads to a higher level of industrialization.With the increasingly mature and popularization of robot technology,there has been an upsurge of industrial robots replacing human tasks.However,the objects that need to be processed are randomly placed,and it is necessary to put the objects on the preset point for the corresponding machine operation,which obviously restricts the application scope of industrial robots.Aiming at the above problems,equipped with visual sensors for industrial robot by combining the monocular camera with the robot to form a coordinated :1531eye system can make the industrial robot automatically complete the positioning and assembly tasks of the target,which can enhance the ability of flexible operation of robots in industrial production.First,the camera imaging system is modeled,and the Zhang Zhengyou camera calibration method is theoretically deduced;Secondly,the problem of hand eye calibration for industrial robot is summarized,and the fast hand eye calibration method used in this paper solves the problem of traditional hand eye calibration method.Finally,through the visual method,the algorithm of position and orientation of target with circular feature is deduced,which can calculate the 3D pose information of objects from two-dimensional images captured by camera.In order to solve the duality problem in pose estimation of circle feature,the method adopted in this paper can eliminate the false solution with the constraint that the plane normal vector determined by the rectangle is parallel to the true normal vector passing over the center of the circle.Through building an experimental platform,the method of target pose measurement based on robot hand-eye vision is validated.The experimental results show that there are approximately parallel relationship between the real normal vector determined by the plane and the normal vector passing over the center of the circle,which the position and orientation of the target can be uniquely determined.It shows that the method of target pose measurement based on robot hand-eye vision is effective.