Research On Vision Location And Measurement Technology Of Industrial Robot

The vision system of an industrial robot is an important system that guides the robot to complete complex tasks.This article focuses on the research of the vision system,builds a three-dimensional coordinate measurement system,completes the parameter calibration of the vision system,and designs an algorithm for the vision system.The principle of the vision system and the effect of actual operation.The design of the vision system mainly refers to the position-based visual servo system.The three-dimensional coordinate measurement algorithm is functionally divided into two parts,two-dimensional coordinate measurement and depth information measurement,which are implemented separately and run independently.The two-dimensional positioning function uses the C++ language to complete the program design,mainly through the monocular camera to capture the image,with the help of image processing technology,the image segmentation of the target in the field of view and the pixel coordinates of the center are obtained.Depth information measurement is realized by fringe projection measurement technology,and the construction of structured light measurement system is included in the vision measurement system.Through this set of vision system,the depth information of the target object is measured.During the measurement process,we noticed the widespread phase error problem in fringe projection measurement technology.Herein,a phase error restraining method based on mean filter was proposed to address the issue of the periodic errors on unwrapped phase,which is caused by gamma nonlinear effect of projectors and cameras.Firstly,the periodic value of phase error is calculated by Fourier transform of unwrapped phase.Secondly,it is designed that a variable length one-dimensional mean filter template,the length of the template is related to the period of phase error.Finally,the unwrapped phase is filtered by mean filter to eliminate periodic phase error.It is showed that the proposed method can reduce periodic error of unwrapped phase by 94%.Furthermore,the distortion ripple is eliminated of flat object's depth information measurement with structured light technology.The method has neither complex mathematical model nor adding fringe pattern,it can improve the accuracy of phase measurement and depth information measurement.Thus,it is more suitable for practical application.