Research On Hand-eye Calibration Technology Of Camellia Fruit Picking Robot

Camellia oleifera fruits have high edible and industrial value,and the Camellia oleifera industry has a very broad prospect in the forestry economy industries.In the current production of Camellia oleifera fruits,the proportion of artificial picking is very high,and the high picking cost and extremely low labor efficiency have seriously restricted the development of the Camellia oleifera industry.In view of the current dilemma of the Camellia oleifera industry,the hand-eye calibration technology used in Camellia fruit picking robot was explored and studied.In this paper,a method of solving hand-eye posture by constructing a hand-eye equation based on the constraints between end-effector motion and camera motion was presented.Based on the existing kinematics researches of the Camellia fruit picking robot,three researches involved in hand-eye calibration technology of Camellia fruit picking robot were studied.A calibration experiment was carried out on a monocular camera of the camellia fruit picking robot.The camera calibration algorithm based on the homography relationship between the image plane and the checkboard plane is used to solve the camera parameters.Based on OpenCV and C++,a camera calibration program is written.The camera's intrinsic matrix,distortion parameters and extrinsic matrices were obtained through the calibration experiment,and the average reprojection error of checkboard corner is only 0.6129 pixels with the experiment result.A hand-eye calibration algorithm based on quaternion was proposed.In this algorithm,the hand-eye equation is rewritten from the matrix form to the unit quaternion form,and the hand-eye equation is solved by a linear optimization.A simulation experiment used to verify the effect of hand-eye calibration algorithm is designed.A hand-eye calibration algorithm based on matrix Kronecker-product is adopted as the comparison algorithm in the simulation experiment.The simulation results show that the proposed algorithm is more accurate and robust than the comparison algorithm,and the average time consumption of the proposed algorithm is only 12ms.An automatic hand-eye calibration system of the Camellia fruit picking robot was designed.The system can automatically acquire calibration data and execute the hand-eye calibration algorithm.Also,the system can limit the movement space of the end-effector with the camera's field of view to ensure that the checkboard can appear completely in the camera's view with great probability,which can bring great improvement in hand-eye calibration efficiency.When the hand-eye calibration result reaches the preset accuracy,the system can automatically terminate.A feasibility test of the system was carried out on the Camellia fruit picking robot experiment platform.The system completes the hand-eye calibration in only 15 minutes,the success rate of the end-effector planning position and orientation reaches 86.5%,and the errors of hand-eye orientation and position calibrated by the system are only 0.0616 and 12.8810mm respectively.The experimental results indicate that the system can accurately,quickly and conveniently complete the hand-eye calibration task of Camellia fruit picking robot.