Design And Implementation Of Intelligent System On Production Process And Personnel Monitoring

With the development of Internet technology,more and more enterprises are opening up the reform of digital and intelligent management mode.The imbalance between the development of informationization and intellectualization and the development of management has gradually emerged.It is of far-reaching significance to assist enterprises from the exploration stage to the practice stage.In this paper,an intelligent management system with high coverage and reusability for most enterprises in this direction is designed for chip production enterprises,which is expected to help the benign development of small and medium-sized entity enterprises in China.Large production workload,complex production process,intensive warehousing and storage of semi-finished products in production,pipeline production mode among process segments,fine division of work among process segments are one of the characteristics of the general chip production process,and slightly omitted production quality will be seriously affected by the type of equipment used in each link,the order of use,etc.Therefore,it is very important to use the intelligent production process management module for production.The production process and personnel monitoring intelligent system designed in this project uses the mainstream B/S architecture to set up the system and complete the business needs such as production management,authority management,document management,etc.At the same time,the system embeds the computer vision module,and uses the target detection and face recognition algorithm to realize the automatic monitoring of personnel.This paper uses cloud-edge collaborative architecture to optimize face recognition algorithm,resolves the pressure brought by traditional cloud center architecture,and effectively reduces response time and transmission costs.The main work of this paper is the design and implementation of the intelligent system for production process and personnel monitoring.The main research work is divided into three parts,which are the front-end architecture development of the intelligent production process management system,the research of the personnel monitoring algorithm based on face recognition and target detection,and the optimization of the cloud-edge collaborative architecture for face recognition.The main work of this paper is as follows:First,front-end and back-end development of WEB intelligent production process management system based on B/S architecture.Firstly,the overview of production process management system from requirement analysis,system architecture and project implementation is introduced.Then,a solution is put forward based on the pain point of chip production management system,which leads to three modules of this systemproduction management,authority management and document management.In each module study,the research process and implementation method of production process management system are detailed from business logic,construction scheme,data table structure,back-end implementation and user interface.This paper establishes a lightweight Web framework using Spring+SpringMVC+MyBatis to build production management system.The backend of the Intelligent System Architecture uses the framework of Spring Boot to process web page requests and read data.The front end establishes logical control of HTML pages and JavaScript pages based on Bootstrap;The database uses MySQL to add indexes,interact with data,and so on,through MyBatis.This management system is implemented by using mainstream and open source development components,so it has the advantages of high reusability,easy expandability,good maintainability,and is conducive to other enterprises to share the results of intelligently digitized management system.Second.Research on personnel monitoring algorithm based on face recognition and target detection.First,the system requirements for personnel monitoring are introduced,that is,assistant attendance and personnel authentication verification.Then,the face detection and face recognition algorithms mentioned in this paper are introduced.Then,the face recognition algorithm used in this system is introduced in detail,including the overview of face recognition technology,system application value and the implementation of specific system face recognition technology.This paper makes a detailed analysis on the algorithm and research of the intelligent system personnel monitoring,which is mainly divided into face detection module and face recognition module.The data set preparation,model training and implementation,test and result verification of each module are introduced one by one.The classifier used in the face detection module is Opencv Github open source XML format classifier.The face detection results of the three Haar feature training cascade classifiers,XML,are verified and analyzed.Finally,the parameters scaleFactor=1.05 and minNeighbors=10 are selected according to the performance indicators.Haarcascade_frontalface_alt as the algorithm model for face detection,has an accuracy of 90.45%,an accuracy of 91.67%,and a recall rate of 72.64%.The face recognition module uses the LBPH algorithm,and the average accuracy is 83.3%,the accuracy is 100%,and the recall rate is 53.9%after ten rounds of performance testing.The recognition accuracy meets the actual scene needs after actual verification and reference.Third.Cloud-edge Collaboration Architecture to Optimize Face Recognition.Firstly,the drawbacks of traditional cloud Service architecture are introduced.And three main features of this system,face detection is unrelated to user data,face detection success is a prerequisite for face recognition,and redundant information can be filtered on the premise of successful face detection,are explained.Therefore,it is naturally supported and highly suitable for cloud-edge collaborative design.This paper introduces the face recognition optimization based on cloudedge collaborative architecture.Comparing this architecture with the traditional scheme optimization,mainly including the comparison with the traditional single application architecture,the comparison with the traditional cloud Service architecture,the optimization of the system’s cloud-edge transport network traffic,and so on.In addition,the edge implementation based on track.js and the cloud implementation based on Haar algorithm are described in detail.Finally,the performance of the optimized result is tested.The optimized percentage of network request content length is 90.21%,the optimized percentage of total network transmission time is 69.06%,and Services such as face detection and image preprocessing can be sinked to the edge to execute.Therefore,face detection and image preprocessing Services will be carried out at the edge,avoiding the performance load,network bandwidth and other pressures caused by the expanding data size on the traditional cloud center architecture,making full use of the computing power of each end,while ensuring information security and data privacy.