Research On Intelligent Detection System For QR-CODE In Industrial Environment

China has been the world's largest manufacturing base,the annual gross industrial production in China accounts for about half of the total GDP.The rapid development of modern advanced manufacturing makes the traditional mode of production unsuitable,the transformation of manufacturing industry is imminent.As the basic industry information carrier,the detection technology for two-dimensional code is the most important in being.Traditional two-dimensional code detection techniques use the way of identification signals trigger point,which has some drawbacks.Is it possible to design a smart detection system to solve these problems? This thesis tries to use object detection,visual tracking and other advanced technology to replace the traditional means of identification signals triggered.In doing so,we can reduce manufacturing costs and improve production efficiency.Visual tracking in industrial production has been widely used,mainly in the automotive,mechanical engineering,electronics,welding and other industries.This thesis takes products with two-dimensional code on a conveyor belt as a motion detection target,and develops a two-dimensional code intelligent detection system.First,we perform target detection and locked using the background subtraction Gaussian mixture model,and extract the two-dimensional code contour on the conveyor belt using the projection and location method.The goal is to grab the two-dimensional code dynamically real-time without trigger signals.However,in the actual industrial production there exist many interference factors,such as more light and vibration.In addition,the product location and size are some differences.All these make a low detection rate for two-dimensional code.This thesis uses the-state-of-the-art matching algorithm to eliminate errors for two-dimensional code in industrial environments.The image processing methods used in this thesis include filtering,denosing,edge detection,positioning correction and others.Studies show that our system can effectively solve the issues of occlusion and inaccurate positioning,and can meet the general requirements for two-dimensional code identification in industry environment line.