The Study On Mechanism Analysis And Locating System With Binocular Cameras Of Tomato Harvesting Manipulator

Agricultural robot is the concentrated embodiment of precision agriculture, micro-electronics, computer and auto-control technique applied in agricultural production. It is an important symbol of automatizing and intelligentizing for agricultural production process.Harvesting is an important step in the process of agricultural yielding process. In order to enhance the productivity and improve the quality, to reduce the laboring intensity, to improve working environment, to realize mechanized ,automatized and intelligentized harvesting, the harvesting robot based on biologic characteristic and planting mode of fruit is developing. At present, some countries, like Japan, America, Holland, have developed several kinds of harvesting robots, which are applying in harvesting vegetable and fruit, i.e., tomatoes, cucumber, strawberry, grape, watermelon, melon, apple, orange, broccoli, etc.. This study has schemed out the manipulator with 7DOF for tomato harvesting, and researched on the kinematic equation and locating system with binocular cameras.A 7DOF manipulator was adopted for tomato harvesting according to the biologic characteristics of tomato, cultivated mode and mechanism type-synthesis regulation of the tomato harvesting manipulator. It took the working place and the dimension of bars as target, built a synthesized objective function of optimization, and obtained the optimal solution of the dimension parameter of tomato harvesting manipulator. In addition, it confirmed the structural dimension according to real work style of tomato harvesting manipulator.After confirming of the manipulator's dimension and structure, the Denavit-Hartenberg method is taken to set the coordinate of tomato harvesting manipulator, and built the Kinematical and Inverse Kinematical model by homogeneous transformation. It also obtained the jacobian matrix of tomato harvesting manipulator. And it did simulation by MATLAB.The kinematics model of tomato harvesting manipulator can make it achieve the requirements for picking fruits, but the position of target must be gained by robot's vision system, while the vision system of tomato harvesting manipulator is to ascertain the target's position. Thus, this study adopted binocular vision system, and built locating model with binocular cameras. It also calibrated the inner and outer parameter of cameras, and did the experiment.