| At present,China is still a big agricultural country,and there are many problems in the process of agricultural development.Among them,manual picking is still the main form of crop harvesting.It not only has high labor intensity,but also has low efficiency due to complex operation,which greatly limits the development of agriculture.Therefore,relying on the rapid development of computer technology and robot technology,designing and developing agricultural robots suitable for agricultural picking environment,replacing the costly manual picking form with intelligent and efficient agricultural robot picking form,will not only play an important role in promoting the development of agriculture,but also be an inevitable trend in the future development of agriculture.In this paper,cucumber was selected as the object of picking.According to the picking environment of cucumber,a cucumber picking manipulator prototype with four degrees of freedom was designed.It mainly consists of three parts: mechanical arm,end-effector and machine vision system.For a start,the structure and control part of the mechanical arm and the end-effector are optimized.The structure of the mechanical arm is determined as the joint coordinate type,using electronic drive and gear drive,and the control system of the mechanical arm is controlled by PCI bus.Based on the bionics idea,a flexible clamping end-effector with multiple modules such as monoblock,clamping,telescoping,cutting and detecting is designed.The picking operation of the end-effector can be realized by controlling the operation of the digital servo and telescopic link with a singlechip computer.In order to ensure that the designed cucumber picking manipulator will not appear structural damage and system instability in practical work,this paper carries out three-dimensional modeling of the cucumber picking manipulator,using ANSYS software and ADAMS software to make a simulation analysis on the mechanical arm and the end-effector respectively.Before that,this paper analyzed the data of cucumber parameters to make the simulation analysis furthest approach the actual cucumber picking conditions.The inherent parameters of cucumber and cucumber fruit stalk,such as average weight,length and diameter,were obtained by statistics.The average friction coefficient of cucumber surface measured by digital inclinometer was 1.62 + 0.06.The minimum positive pressure of cucumber skin damage and the maximum shear force of cucumber fruit stalk separation were 4.63 N and 5.35 N respectively by WDW-20 universal tension and compression testing machine.After obtaining the physical parameters of the cucumber,the static analysis,modal analysis and dynamic analysis of the mechanical arm and end-effector are carried out in this paper.In the static analysis,the maximum deformation of the finger occurs at the fingertip,and the maximum stress occurs at the joint between the rubber finger surface and the elastic finger skeleton steel sheet,which is about 167.64 MPa,far less than the yield strength of the material(784 MPa).This meets the design requirements.In modal analysis,the maximum vibration amplitude of the mechanical arm made of bakelite plate and aluminium alloy in the sixth mode is 0.24 mm,which has smaller vibration amplitude and better stability,and is more in line with the design requirements of the mechanical arm.In dynamic analysis,the end-effector can provide an appropriate clamping force to complete the non-destructive picking of cucumber.There is no significant sudden change in the moment,angular velocity,angular acceleration and displacement of the four joints of the manipulator,which meets the requirements of actual picking.In the design of the vision system,the Kinect sensor is selected as the camera fixed on the mounting bracket,and the real-time pictures data of manipulator operation are obtained by moving along the waist joint.The environment configuration of the Kinect for Windows SDK and Opencv-3.0.0 database is completed under the Microsoft Visual Studio 2010 platform.Based on the HSV color model,the acquired color image is processed to realize the recognition of cucumber,and the pixel coordinates of the minimum circumscribed rectangle and the minimum circumscribed rectangular heart point of cucumber are obtained.The color image is mapped to the depth image and registration is completed.The world space coordinates of the feature points in the camera coordinate system are obtained through the transformation relationship between the inherent coordinate systems inside the Kinect sensor,and the cucumber location is realized.Through the error test obtained by the coordinate values of the feature points,the error between the obtained coordinate values and the measured coordinate values is generally normal distribution in the three directions of X,Y and Z.The error in the three directions is mainly concentrated at 2 ~5 mm.Within the allowable range of error,the vision system can complete the entire image processing in only 5.3 s,which has a better real-time visual recognition.After completing the above work,the motion control test of the mechanical arm and end-effector and the simulation picking test of the manipulator have been completed successively.Through repeated simulation tests show that the average time required for vision system to complete the recognition and localization of visual system and for manipulator to complete picking and releasing was 20.14 s and the success rates of localization and picking were 92% and 84% respectively.The experimental results show that the motion and control systems of each joint of the mechanical arm cooperate well,the hypogynous machine and the industrial control machine realize good signal communication,and the vision system cooperates well with each joint of the manipulator and has high recognition and positioning accuracy.The optimized end-effector is stable and flexible,and can complete the picking operation without damaging the cucumber epidermis. |
PDF Full Text Request