Research On Key Technology Of Visual Guide Flange Assembly

With the development of industrialization,industrial robots are becoming more and more intelligent.The application of machine vision technology in industrial robots is more and more.At present,the process of machine vision replacing human is speeding up.The application of vision guidance technology in industrial robots is more and more.In this paper,taking flange plate as an example,the reinforcement to be assembled is fixed,and the manipulator is guided to move the flange plate to the exact position by continuously collecting the flange plate image with the camera,so as to realize the final assembly.This paper first introduces the current development of assembly robot,analyzes the current situation,points out the current problems and future development trend.Then,the scheme design and the selection of software and hardware are carried out for the assembly of flange plate,and then the camera selected is calibrated by Zhang Zhengyou calibration method.Then according to the needs of assembly,a method of pixel size calculation based on facula is proposed.This method can not only calculate the pixel size,but also provide help for flange tilt correction.Secondly,the image of flange plate is collected and preprocessed.Image preprocessing mainly includes image filtering,image segmentation,morphological processing,edge detection and so on.The median filter is selected to suppress the inherent noise in the image,and the best threshold is selected for image segmentation through continuous experiments.In order to remove holes and edge noise,morphological operation is selected for processing.The Canny operator is selected as the edge detection algorithm of the system.Then calculate the roundness and select an assembly hole to locate the center of the circle below.In order to ensure the accuracy of assembly,the concept of sub-pixel edge detection is proposed based on the traditional edge detection results.The Zernike moment method is selected based on the three subpixel edge detection methods,namely,the contrast moment method,the fitting method and the interpolation method.After that,the inclination angle and rotation angle are calculated by using the relationship between circle and ellipse to realize the inclination correction of flange.Finally,the accurate least square method is used to locate the center of the flange after the tilt correction.The displacement with the accurate position and the rotation angle of the assembly hole are calculated by coordinate transformation.Finally,vs2017 and MFC are used to write the system interface program,and the system algorithm is analyzed and verified by experiments.The experimental results show that the assembly system can output data within 0.4s,and the accuracy can reach 0.05 mm.