In view of the current problems such as low positioning accuracy,restricting production efficiency and potential safety in manually guiding crane tube docking during oil loading,an automatic crane tube system based on machine vision and positioning was proposed.Based on the traditional crane tube joints,automation reconstruction was completed and a balancing mechanism was designed to coordinate the weight of the arm and the vertical tube; based on a simplified three-dimensional model,statics analysis was carried out;the kinematics model of automatic crane tube was established,and the workspace analysis and trajectory planning were carried out;the program design of tank mouth identification and location was completed by machine vision technology.The image simulation locating experimental results show that the locating error of the visual system is within 8 mm,it can meet the locating and docking requirements of the crane tube.