The Design And Implementation Of Tire Defect Real-time Monitoring System

In the field of tire industrial production,defect detection is an important link of tire quality assurance.The defect detection method based on manual observation of X-ray image is timeconsuming and labor-consuming.Therefore,this paper designs and implements a real-time tire defect monitoring system.The system is an industrial control system combining industrial hardware equipment and digital image processing technology.It is used in the quality detection link of tire production process,which is conducive to improving the accuracy of tire defect detection,reducing time cost and promoting industrial production capacity.Firstly,this paper summarizes the background and non system problems of tire production from the perspective of industrial demand and non system.Then it puts forward a detailed technical scheme,implementation scheme and test scheme,and constructs the system in strict accordance with the above scheme.In the research of image defect recognition algorithm,this paper constructs a general defect detection process based on a large number of literature at home and abroad,combined with machine learning methods and traditional image recognition methods.From a technical point of view,the system is divided into three parts:client,server and algorithm subsystem.The client and server communicate based on websocket and HTTP.The server starts the sub process to call the algorithm subsystem and interacts with the algorithm subsystem according to the parameter specification.This paper uses C++and QT framework to build the client program,which has high efficiency and can interact well with external devices;Adopt node JS build the server program and use nginx as websocket and HTTP proxy server to reduce the deployment and maintenance cost and greatly improve the stability and availability of the system;Redis storage system global parameters and action instructions are adopted;The express framework is used to process HTTP requests and store the data generated in the system operation in mongodb.For the algorithm subsystem,this paper adopts Python combined with scikit learn library training model,and compiles the algorithm subsystem program into DLL when deploying the system.In addition,this paper proposes and designs in detail the core mechanisms in the operation of the system,such as locking mechanism,action instruction mechanism and message frame encapsulation mechanism.These mechanisms play a vital role in ensuring the correct execution of detection tasks.From the perspective of function,the system is divided into five modules:image transmission,client core module,server core module,parameter configuration and data management.This paper describes the operation process of each module in detail.From the perspective of algorithm implementation,this paper realizes the judgment task of tire inner and outer sides based on SVM,and realizes the tire defect detection task based on sidewall debris recognition algorithm and texture independent crown crack recognition algorithm.These algorithms have high recognition accuracy and excellent performance,and have achieved good results in application.After the system is deployed to the production environment,it can run continuously and stably,perform well in accuracy and real-time,and can meet the needs of industrial production.The research of tire defect real-time monitoring system is of positive significance,in line with the digital and intelligent development goal of manufacturing industry,and can promote the transformation of manufacturing industry to a certain extent and form the industrial ecology of intelligent manufacturing.