Smart Station System Based On Big Data Of Face

With the development of transportation in our country,new features and changes have appeared in the construction and development of China’s automobile terminal.In recent years,with the increasing number of passenger terminals,new requirements have been put forward on the monitoring and management of passenger terminals,safety precautions,accident rescue and dispatching command.However,the traditional video surveillance system is commonly used in the bus terminal in our country.However,due to the huge passenger flow,complex environment and many regulatory requirements,the current portrait recognition system is difficult to meet the needs of the current bus terminal in terms of performance and function.Based on the related concepts of smart city,safe city and big data,this paper takes Panyu Bus Terminal as the application object,and based on the relative concepts of smart city,big city and big data,combined with the portrait recognition and comparison technology,build intelligent,safe and convenient intelligent bus terminal comprehensively for the management of the bus terminal Management system.Panyu Bus Terminal is the hub of Guangzhou’s largest road passenger transport and the major transport hub of highway passenger transport in southern China.In view of Panyu Bus Terminal daily responsible for the transport of a large number of passenger flow,how to enhance the terminal level of information,how to improve the safety level of the bus terminal,to create a safe,convenient and intelligent terminal is Panyu Bus Terminal managers and government management Department of the current important work.The portrait recognition system is the core part of the smart terminal system.In view of the highly non-restrictive environment characteristics of the terminal,a Seeta Face portrait recognition framework is proposed.It is proved that SeetaFace has more features on LFW datasets than the LFW and SCface datasets Effective performance.This paper solves the technical problems of portraits identification in two video surveillance stations.Firstly,it analyzes the influence of passenger light changes in portrait recognition,analyzes the method of weighted variation model,and carries out under daylight and night vision scenarios Experiments were carried out to improve the recognition rate of deep learning model by using illumination preprocessing enhancement method.Secondly,this paper analyzes the influence of passenger attitude of passenger terminal on portrait recognition and proposes a method of attitude pre-assessment.It also verifies that the pose pre-assessment method proposed in this paper has high recognition efficiency.After solving the technical problem of the portrait recognition of the terminal,the video surveillance portrait recognition system of the terminal based on SeetaFace is designed and implemented.With this core module,an intelligent,safe and convenient smart terminal is developed for the safety management of the Panyu Terminal.Integrated Management System.While solving the problems of low processing capacity and slow response speed of the existing portrait recognition system,taking the Panyu Terminal as an example,the actual operation and function of the intelligent terminal system are tested to ensure the actual operation and management of the intelligent terminal system effect.