| With the developments of motor technology,the requirement of leakage current produced during commutation is increasing. As one of the key factors, dimensional ac-curacy of mica slot in commutator greatly limits the application of brush DC motors. Thus, to improve the positioning accuracy and speeds on the processing of milling mica slot, making a study on visual position technology is very significant for mass produc-tion of the commutator and reducing production cost.To meet the requirements of high positioning accuracy (0.05mm) and high posi-tioning speed (0.5s / line), and develop a set of slotter system based on visual position, This paper made a systematical study on visual recognition, calibration and servo tech-nology involved in visual positioning.Firstly, analyzed median filter and mean filter, then took the characteristics of the target image into account, proposed a hybrid image filtering algorithm based on the shape characteristics of the mica slot.Secondly, compared the advantages between the algorithms of traditional image segmentation, then proposed an adaptive thresholding segmentation algorithm based on edge detection in order to combine the accuracy of edge detection with the efficiency of threshold segmentation.According to the characteristics of positioning system in the project (central visual position), established linear imaging model for it, then completed camera calibration by perspective transformation method. Meanwhile, in order to meet the requirements of high accuracy and high positioning speed, constructed visual servo system based on the stationary image jacobian matrix, completed simulation on the position algorithm.Finally, after demonstrating the feasibility of positioning algorithm, designed the hardware platform of positioning system, and completed the procedures. Did statistical analysis of positioning time and positioning accuracy during the experiment. |
PDF Full Text Request