欢迎访问铁路通信信号工程技术,今天是 English

铁路通信信号工程技术 ›› 2025, Vol. 22 ›› Issue (2): 91-97.DOI: 10.3969/j.issn.1673-4440.2025.02.014

• • 上一篇    下一篇

城市轨道交通人脸识别算法配置方案

张炳森1,乔宁宁2,付保明3,张 宁4   

  1. 1.北京城建设计发展集团股份有限公司 北京 100045;
    2.北京市地铁运营有限公司 北京 100044;
    3.苏州轨道交通建设有限公司,江苏苏州 215004;
    4.东南大学智能运输系统研究中心轨道交通研究所,南京 210008
  • 收稿日期:2024-10-23 修回日期:2025-01-17 出版日期:2025-02-25 发布日期:2025-02-25
  • 作者简介:张炳森(1992—),男,工程师,硕士,主要研究方向:轨道交通弱电系统设计,邮箱:604292994@qq.com。
  • 基金资助:
    国家重点研发计划资助项目(2020YFB1600700)

Configuration Scheme of Facial Recognition Algorithm for Urban Rail

Zhang Bingsen1,  Qiao Ningning2,  Fu Baoming3,  Zhang Ning4   

  1. 1.Beijing Urban Construction Design & Development Group Co., Ltd., Beijing    100045, China;
    2. Beijing Metro Operation Co., Ltd., Beijing    100044, China;
    3. Suzhou Rail Transit Construction Co., Ltd., Suzhou    215004, China;
    4. ITS Rail Transit Research Institute of Southeast University, Nanjing    210018, China
  • Received:2024-10-23 Revised:2025-01-17 Online:2025-02-25 Published:2025-02-25

摘要: 人脸识别算法是人脸识别应用核心,需对其配置方法进行系统性分析,从而为人脸识别系统建设提供一套科学指导方案。介绍人脸识别平台系统架构,分析人脸识别应用流程、业务管理平台和人脸算法平台功能定位,明确人脸算法重要性;结合轨道交通票务系统业务需求以及人脸识别技术特点,提出票务系统中人脸识别算法3种配置方案:单算法配置、双算法主备配置、双算法双活配置;从建设及维护成本、可靠性、可扩展性等方面对各方案进行综合性对比,分析不同规模城市方案适应性选择,并针对不同方案弊端给出优化建议,为城市轨道交通人脸识别系统建设提供一定参考。

关键词: 城市轨道交通, 票务系统, 人脸识别, 配置方案

Abstract: The facial recognition algorithm is the core of facial recognition applications, and its configuration methods need to be systematically analyzed, to provide a scientific guidance plan for the construction of facial recognition systems. Firstly, this paper introduces the system architecture of the facial recognition platform, analyzes the process of facial recognition applications, the functional positioning of the service management platform and the facial algorithm platform, and clarifies the importance of facial algorithms. Then, in view of the service requirements of the rail transit ticketing system and the technical characteristics of facial recognition, it proposes three configuration schemes for facial recognition algorithms in the ticketing system: single algorithm configuration, dual algorithm configuration in main and backup mode, and dual algorithm configuration in dual active mode. Finally, it carries out a comprehensive comparison of various schemes from the aspects of construction and maintenance costs, reliability, scalability, etc., analyzes the scheme adaptability for cities with different scales, and proposes optimization suggestions for the drawbacks of different schemes. The research on facial algorithm configuration schemes can provide certain reference for the construction of facial recognition systems for urban rail transit.

Key words: urban rail transit, ticketing system, facial recognition, configuration scheme

中图分类号: