Research On A Machine Vision Based Graft Clip Regulating Device

The current state of automation of the clamping operation for the existing automated grafting machine is not high.Based on computer vision technology and mechanical arm,this paper studies a grafting clip finishing device,which provides the basis for the miniaturization and automation of the grafting machine.The innovations of this paper are as follows:(1)A method for identifying the location and state of grafting clips is proposed.(2)In order to achieve flexible gripping and finishing,a terminal posture self-retaining robot arm with large motion range is designed.The research work of this paper mainly has the following aspects:(1)Research on the method of positioning and identification of grafting clips based on machine vision.In order to realize the precise positioning of the gripping point and the rotation angle of the grafting clip during the automatic clamping of the melon and solanaceous crops.The target image acquired by the machine vision is processed,and the contour chain corner point and the centroid are identified according to the extracted target contour chain.Through a series of calculations,the grab point and the angle of rotation are obtained.The results show that the average point of the grabbing point of the grafting clip positioning recognition method based on machine vision is 2.1 pixels.(2)Research on SVM-based target object pose classification method.For the target recognition object in the actual working environment,it may be upright,tilted,and adhered to other target objects.In this paper,the SVM-based classification method is used to classify the target pose,and select the centroid and the farthest pixel distance.The feature k of the ratio of the area ratio of the contour chain region,the feature m describing the number of the corner points of the contour chain,and the feature n describing the number of concave corner points of the contour chain,the three characteristic parameters are modeled as input parameters of the SVM training.The average correct rate of classification is 97.56%.The measurement results of different angles of different angles indicate that when the working height of the camera is 10 cm to 20 cm,when the long axis radius centered on the camera is 210 pixels and the short axis radius is 156 pixels,each.The attitude classification effect is above 93%,which is the best classification area.(3)A terminal attitude self-holding robot arm design.For the existing small mechanical arm part adopts four-bar linkage mechanism,there is terminal self-holding scheme but the range of motion is limited;some adopt synchronous belt mechanism transmission power,and there is no terminal attitude self-holding scheme.A mechanical arm that realizes the self-holding of the terminal posture under the premise of ensuring the range of motion of the large and small arms is designed.The mechanical arm can ensure the self-holding of the terminal posture and the range of the arm movement under the premise of ensuring the range of motion of the large and small arms.0 ° ~ 320 °,boom range 0 ° ~ 270 °.(4)The motion control and implementation of the grafting clip finishing system.In this paper,the D-H method is used to analyze the kinematics of the manipulator,and the analytical solution of the forward and inverse kinematics is obtained.Based on the forward and backward kinematics,the accuracy analysis of the mechanical arm repeat positioning is carried out,and the operation mode of selecting the maximum speed of 90% and performing the error clearing operation every 5 working cycles is determined.A grafting clip grabbing terminal was designed for the grafting clips of the platagram and the round mouth.The maximum opening of the terminal was mm,and the clamping height was kept constant during the clamping process.Finally,this paper combines the above research content and stepper motor control system to realize a grafting clip active finishing device based on machine vision.