Design And Implementation Of Big Data Technology Based Access Control Traffic Data Analysis Platform

With the advent of the age of information and intelligence,access control systems based on 3D face recognition have been widely used in colleges and universities as a core application of digital security.This kind of access control system can reduce the threat from outside the campus,thus safeguarding the safety of people and property inside the campus.However,the existing access control systems have significant shortcomings in terms of storage,management and analysis of the massive amount of data.How to properly handle and use these data has become an important issue of concern for schools.In this context,big data technologies have emerged and are widely recognized and used in academia as well as industry.Among them,Hadoop technology has become the mainstream platform in the field of big data processing with its good ecosystem and the advantages of high scalability,stability and fault tolerance.Open source data warehouses based on Hadoop clusters can assist schools in management.Therefore,more and more universities are building their own data warehouse systems to promote the development of school informatization to a new level.Data warehouses have a topic-oriented approach and can process,handle and integrate data sets that change over time.However,traditional data warehouses can only handle data that is partially structured and cannot handle other types of data.The data on campus,on the other hand,is not only structured data,but also unstructured data as well as semi-structured data.For other unstructured data,data warehouses are not able to manage it well;nor can they solve the "data silos" well.In order to solve these problems and break the "data silos",the potential value of the data is brought into play.Combined with the business characteristics of campus data and business needs,it is of great practical significance to design and implement a data analysis platform for the storage,management,analysis and statistical mining of access control data based on 3D face recognition.The main research work of this thesis is as follows:1.Analyze the business needs of the access control data analysis platform based on Big Data Technology,including data backup,use data lakes for data storage and management,data desensitization,design and implementation of data warehouses in data lakes,use knowledge graphs for data visualization display,automated task processes,data analysis,mining statistics,background management systems,and other related modules.2.Applying big data technologies such as data lakes to the storage,management and analysis of campus data.while using Kimball dimensional modelling theory combined with data hierarchy to design a data warehouse based on data lakes,and using Apache Hive to complete the data warehouse construction.In conjunction with the Azkaban task scheduling tool,it is designed to automate the processing of newly added data on a daily basis.3.Design a detection method for abnormal behavior of personnel based on campus access control data.The Prefix Span algorithm is used to pattern mine path sequences as well as time sequences constructed from access control users and their daily entry and exit locations and times to generate a library of normal sequences.The abnormal behaviour detection is completed by quantitatively portraying the previous day’s behaviour sequences using relative edit distance as well as relative support.Through experimental analysis,the LSTM prediction model designed in this paper was used on the chosen dataset,and the results were better than traditional models such as ARIMA.Meanwhile,according to the set rules,the users were counted out and not returned.4.Use Neo4 j to build a knowledge graph for relevant structured data in the data lake,and display nodes and the relationships between nodes by using points and edges.5.Design and implementation of a backend management system to manage the platform.Use technologies such as SSM + Spring Boot + VUE as the framework for the system to manage the platform users and user permissions.Finally test the whole platform to verify the usability of the platform.