3D Reconstruction Technology Of Blueberry Fruit And Its Discrete Element Parameter Calibration

When the blueberry picking machine is in operation,there is a large amount of blueberry fruit movement.The movement of the blueberry fruit is not only related to its own mechanical properties,but also related to the structural form and size parameters of the working parts of the related blueberry picking machine.A good blueberry picking machine design should make the blueberry fruit move according to the expected program.For example,after the blueberry fruit is detached from the stalk by the vibration of the picker’s excitation mechanism,the collision between the blueberry and the picking machine body during the subsequent fruit movement should be reduced and reduce the damage of blueberries.Discrete Element Eethod(DEM)is a universal and effective method for studying the dynamics of granular bodies.It can calculate and simulate the movement,collection and transportation of blueberry fruits in the picking machine after they fall off the stalk.When using the discrete element method to calculate and simulate the blueberry picking process,it is necessary to construct a blueberry fruit model,among which the creation of the three-dimensional structure model of the blueberry fruit and the discrete element parameter setting of the blueberry model are indispensable and important parts.The structure of the blueberry fruit discrete element model can be obtained through the three-dimensional reconstruction of the blueberry fruit.The structure from motion(SFM)algorithm of computer vision is used to reconstruct the blueberry fruit.The SFM algorithm matches a set of sequence images.The calculation of feature points can effectively restore the three-dimensional morphological structure of the blueberry fruit in the image.The discrete element parameters of the blueberry fruit model can be indirectly calibrated and optimized through a series of experiments,and the reliability of the obtained parameters can be verified through verification experiments.In this paper,a monocular camera is used to study the three-dimensional reconstruction method of blueberry fruit,and the structure from motion algorithm is used to restore the three-dimensional spatial structure of blueberry fruit,and further study the structure parameters of blueberry fruit.The key technologies of the blueberry fruit three-dimensional reconstruction process are studied in depth from the aspects of monocular camera calibration,blueberry fruit image feature point extraction,feature point matching and 3D point cloud generation and structural parameter extraction.Through monocular camera calibration,the camera parameters f,u0,v0,k1,k2,p1 and p2 are calculated to be 21.66mm,959.73,663.56,-9.73e-02,-4.88e-02,-1.89e-04,8.11e-04.Through the comparison of different feature point detection algorithms,the most suitable SIFT algorithm is selected.Through feature point matching,the corresponding point pairs of blueberry fruit images from different angles are obtained.The SFM algorithm is used to reconstruct the three-dimensional point cloud model of blueberry fruit.After point cloud processing,the diameter and height of the blueberry fruit are extracted.The correlation coefficient with the measured value is 0.88,0.84.Aiming at the problem that it is difficult to measure the discrete element parameters of blueberry fruit,this paper takes the angle of repose of the blueberry fruit stacking test as the response value and the discrete element parameter as the independent variable.The Plackett-Burman(PB)Design test,the steepest climbing test,and the Central Composite Design test were carried out in sequence to calibrate and optimize the discrete element parameters of the blueberry fruit.The blueberry-blueberry recovery coefficient was 0.06,and the blueberry-blueberry static friction coefficient was 0.56.The blueberry-steel static friction coefficient is 0.70.The discrete element simulation results are compared with the blueberry accumulation physical test and the relative error is 0.57%.The use of computer vision technology to carry out three-dimensional reconstruction of blueberry fruits provides a certain theoretical research foundation for the subsequent installation of a vision system on the blueberry picking machine to realize automatic and intelligent picking.Through the reconstruction of the blueberry fruit three-dimensional model and the calibration of the discrete element parameters,it has laid a good data foundation for the use of discrete element method to simulate the movement process of blueberries after picking.It is the structural form of the working parts of the blueberry picking machine using the discrete element method,a necessary prerequisite for optimizing size parameters.