Research On The Rapid Identification Of QR Code For Visual Positioning

With the advent of the industry 4.0 era,automation has become more and more important in traditional industries.“Intelligent Manufacturing” and “Machine Substitution” have become the common goals faced by each manufacturer.Vision is the eyes of robots,which play a central role in this industrial revolution.As a mature coding method,two-dimensional code can be used as a powerful complement to VSLAM technology because of its advantages such as visual stability,high efficiency,low cost,and simple production.In recent years,an increasing number of applications have appeared in robot assistant positioning.In the meantime,it has attracted the attention of many scholars and engineers.This paper takes the visual assistant positioning of wheeled robots as the application background,combines the actual requirements of the industrial automation industry,analyzes and studies the algorithms of image preprocessing,image registration,template matching,error correction code encoding and decoding,and designs a suitable two-dimensional code identification solution for the needs of visual assistant positioning of wheeled robots,including frequency domain high-pass filtering,Position Detection Patterns recognition and sorting,code region coarse positioning algorithm based on indirect adjustment,and fine positioning algorithm for code area based on image block gradient,code area design.The two-dimensional code recognition algorithm suitable for the visual assistant positioning of wheeled robots is implemented by C++ and opencv,this module added abstract factory,multi-threaded,double buffering architecture,which makes the decoding of two-dimensional code more robust,more real-time and more efficient.Finally,through the laboratory's debugging and testing of two-dimensional code identification's routine on the machine vision experimental platform(simulating robot site),the effectiveness of the proposed algorithm is verified.The algorithm of this paper is more suitable for the visual assistant positioning of robots than other two-dimensional code recognition algorithms.It makes a good balance between recognition rate,accuracy and software development cost,and can lay the groundwork for the future robots with faster and more accurate positioning.In addition,this paper also has positive reference price to traditional two-dimensional code recognition.