Design And Implementation Of Grabbing Flying Object System Based On Kinect Sensor

Robot grasping based on stereo vision information is a key research direction in robotics field in recent years.It has high application value in the field of industrial production and life service.But,for now,most grasping tasks are aimed at static targets.There are relatively few researches on grabbing flying object.Based on stereo vision information,we research the grasping technology of flying objects.It is significative to expand the application field of robot grasping and improve the technology of grabbing flying object.This paper uses a visual system composed of KINECT sensors and ZED stereo cameras.The grab mechanism consists of the UR5 robots and barretthands.Design and implementation of a grabbing flying object system based on stereo vision information.Through the KINECT's depth sensor,the 3D coordinates of the flight trajectory point are realized.Two prediction models are used to calculate the impact point of the badminton in real time.By calculating the camera transfer matrix and the Eye-to-hand calibration,the grasping point coordinates are transformed into the coordinate values in the coordinate system of the UR5 robot.When we testing a grabbing system which target on flight object,it is hard to get error between predicting impact point and real-world impact point.To solve this problem,an approach is proposed to get error of grabbing based on Depth-image and OpenCV.With this approach,stereo camera can record key frames automatically,by combining the Kalman filter and image tracker,this method could get the position of gripper and target,calculate error and report it on line.Through experimental,we have proved that the grasping system which we designed for grabbing flying objects can achieve the mission goal.