Text / China Minsheng Bank Shanghai Branch, Li Qiang, Li Yanyue, Shang Xiaodong, Gao Jinyu
With the continuous development of science and technology and artificial intelligence, the application of intelligent security in the security management of the vault is also increasing. This article will discuss some personnel management, process links and application scenarios involved in the application of personnel identification, access management, security inspection and intelligent security in the vault.
Promotion and prospecting of contract verification procedures
Analysis of the risk points existing in the security management of bank vaultsBank vaults are mainly used to store a large amount of cash, securities, precious metals and other important financial assets. The security of the vault requires the combination of civil air defense, physical defense, and technical defense, so as to create a solid physical environment and management place.
According to publicly available data, in recent years, there have been three cases of theft of bank vaults in China, such as the theft of 2 million yuan from the 128 vaults of the central bank in 2000, in which the treasury keepers stole a long-term immovable vault at one time; in 2001, 100,000 yuan in cash was stolen from the vault of the Hunan Post Office, and the cash in the vault was stolen for the purpose of escorting and handing over; and in 2007, the 50 million yuan theft of the vault of the Bank of Handan was a case in which the vault keeper cooperated with two people to steal cash from the vault for a long time. From the above cases, it is not difficult to find that the vault technical defense, physical defense facilities are solid and complete, and it is almost impossible for outsiders to destroy the success of the robbery, and the loophole is usually committed by internal personnel, especially the vault staff who use their positions to steal or collude to commit crimes, which has the characteristics of a long time span and strong concealment. Although the bank has formulated detailed rules and regulations and perfect operation specifications for the management of the treasury, is equipped with 24-hour real-time monitoring of personnel, and has adopted regular or irregular physical inventory inspection methods, it is very easy to have visual fatigue, slackness, negligence, and negligence because the work content is single for a long time, the physical inventory is mainly manual, the technical defense facilities do not have intelligent analysis functions, and the staff around them are familiar faces.
It can be seen that the personnel working in the vault are the unstable factors and main risk points in the current treasury management.
The application of intelligent security in the treasury management scenario
With the help of image recognition, biometrics, and machine learning functions of the most popular AI technology, the vault intelligent security can explore and mine the risk control information in massive video surveillance, establish a large model of bank vault risk control, transform video images into data analysis models, learn through the model, screen abnormal data in similar video data, upgrade a large number of manual identification and judgment to AI system automatic screening and screening, timely warning or regular push abnormal reports according to demand, and improve treasury risk monitoring and early warning. The timeliness and continuity of abnormal analysis can realize the intelligent construction of the vault, realize the organic integration of civil air defense and technical defense, and eliminate the uncontrollability caused by human factors.
1. Apply intelligent security technology to identify the identity of personnel entering and exiting the vault
The management of the qualifications of personnel entering and leaving the vault is the source of risk control in the vault. Strict scrutiny can be carried out through the combination of the AI identity verification system with the bank's internal audit system. Personnel entering and leaving the vault are generally divided into three categories: fixed personnel (treasury clerks), business personnel (handover personnel of banks or escort companies), and temporary personnel (professional maintenance and inspection personnel).
Biometrics and Authentication: Install biometric devices (such as facial recognition devices) at the entrance of the vault to authenticate people entering and exiting the vault, and conduct real-time comparison through AI technology to ensure that only authorized personnel can enter the vault.
Access control system: After the comparison is passed, the intelligent access control system can be turned on to ensure that only authorized personnel can enter the relevant areas of the vault. Each employee can be assigned unique access permissions, reducing the risk of intrusion.
Two-factor authentication: Access to core areas such as cash storage areas can employ two-factor authentication, such as fingerprint recognition, iris recognition, or facial recognition, to ensure that only authorized personnel can enter.
Identity tracking and identification: Personnel who enter the core area of the vault after secondary identity verification will be tracked by various cameras during the vault to prevent unauthorized personnel from tailgating into the core area.
2. Apply intelligent security technology to strictly control the workflow of treasury business personnel
In accordance with the five same systems of "same opening, same entry, same presence, same lock, and same exit", the treasury management uses image recognition, target detection, tracking, alarm push and other technologies to intelligently identify and monitor the whole process of personnel, operation behaviors, trajectories, and abnormal situations in the vault according to the six scenarios of incoming, outgoing, detection, trajectory, clearing, and closing, so as to facilitate timely discovery and disposal of abnormal situations.
(1) Storage: Through identity verification and biometric technology, the personnel entering the core area of the vault are verified twice, which can effectively prevent illegal personnel from entering the core area of the vault and improve the security of the vault. At the same time, it is also possible to manage the attendance of employees in the vault and record abnormal situations such as single person entering the warehouse.
Fixed personnel: Enter after taking "access control authority + post status + face recognition" authentication. Select "Reason" on the system, and the three people will pass and enter in turn.
Business personnel: enter after taking "two-way authentication + post status + face recognition". Select "Reason" on the system, and the business personnel will swipe their faces for authentication in turn (the system will automatically verify the identity of the escort company's filing personnel and the identity of the handover personnel on the same day of the bank's filing), and the fixed personnel will enter the handover area after face verification.
Temporary personnel: After the "ID card + face recognition + approval form" is certified, they will enter after being reviewed by the fixed personnel of the vault, and the accompanying personnel will accompany them throughout the process. On the system, find the task approval form pushed by the platform (verify the validity of the bank's internal approvals at all levels, collect the personal information of maintenance personnel in advance in the approval process, and automatically eliminate potential risk personnel with bad behavior records and short working hours), temporary personnel face + ID card verification, accompanying personnel authentication, and fixed personnel face authentication in turn.
(2) Outbound: Intelligently identify whether the number of people and personnel in and out of the warehouse are consistent, whether the number of boxes in and out of the warehouse is consistent, whether the changes in the box are consistent with the business system, and the closing state of the facilities out of the warehouse, such as turning off the lights, etc., for video analysis and abnormal reporting.
Fixed personnel: 3 people in turn to brush their faces, leave the warehouse, and end the task (if there is an accident, they can be forced to exit).
Business personnel: 1 fixed person starts the outbound task, business personnel brush their faces in turn, and fixed personnel brush their faces in turn to end the task.
Temporary personnel: 1 fixed person starts the outbound space, temporary personnel swipe their faces in turn, accompanying personnel scan their faces for authentication, and fixed personnel swipe their faces in turn to end the task.
(3) Detection: There are many cameras in the vault, and AI recognition technology is used to form multiple detection models, and anomaly detection runs through the whole process of the operation of the vault business personnel. For example, the box is blocked by detection, and the stacking of the box is detected in time to cover the camera, and there is a blind spot in the monitoring; the number of people is detected, the number of people present in the vault is less than the number of people in the warehouse; the shift detection, the position of the box in a fixed position for a long time changes; the clearing behavior detection, the abnormal situation such as a single cash out of the clearing area appears in the clearing process. Alerts are issued for abnormal situations.
(4) Trajectory: Intelligently identify information such as the activity trajectory of the treasury keeper, such as whether there is any omission in the daily inspection of the safety facilities and fire protection facilities of the vault, and whether there is any omission in the daily inspection of the vault area. For the number of non-compliance or the long stay time, the system will automatically remind and push the abnormal report to the supervisor at the end of the day.
(5) Clearing: By combining with the monitoring system, it can track the entry and exit and movement of items in real time, record the opening frequency of the storage box, and exceed the data of the normal opening frequency, realize automatic data comparison and checking, and push the report in time when abnormalities are found.
(6) Closure: Before closing, the warehouse keeper needs to check the completion of the day's business events and other events (prompts, alarm reminders, etc.) issued by the system, as well as the management of the site.
3. Apply intelligent security technology to improve the quality of work of vault guards
Using AI technology, the camera in the vault can achieve real-time monitoring and automatic identification of abnormal behavior, and when abnormal situations, suspicious persons, items or abnormal actions are detected, the system will immediately send out an alarm, which can monitor and alarm abnormal events in the vault in real time, and remind the vault guards and managers to take measures.
Supervision of the performance of duties by vault guards. The off-duty detection camera is used to detect the handover of the guards' face-slapping and punching records and on-the-job conditions in real time to confirm their on-duty status, as well as whether there are violations such as sleeping on duty, leaving the post, and looking at mobile phones, so as to ensure that the guards are on duty normally.
Abnormal warning around the vault: Through the intelligent analysis of surveillance video, the abnormal situation around the vault can be analyzed in real time, such as someone trying to approach the vault or carry out suspected sabotage behaviors such as climbing over the wall, and strange vehicles staying for a long time, etc., the system can give early warning to ensure that the vault guards take timely measures.
Automatic O&M of the monitoring system. According to the set plan, it automatically detects and prompts alarms for abnormal states such as video signal loss, blurred images, camera blocking, and disconnection of all surveillance cameras in the vault area, so as to prevent the treasury guards from being unable to find problems in the monitoring equipment in time due to their lack of serious performance of duties.
Other intelligent security technologies improve the quality of security management. For example, the use of intrusion detection and early warning, temperature and humidity warning, sound recognition early warning and other equipment, combined with artificial intelligence technology, when an abnormal situation occurs, the system can automatically trigger an alarm, and at the same time send the detected abnormal situation to the vault guards in time for reminder, attention and disposal.
4. Apply intelligent security technology to improve the security inspection effect of the vault
The inspection of bank vaults usually adopts physical inspections, cash and valuables inventory, security system inspections, entry and exit records, monitoring of key parts and operations, random spot checks, to detect potential irregularities or security vulnerabilities, with the help of intelligent technology can improve the pertinence and efficiency of inspections.
Use AI bots to carry out intelligent patrols. The robot can realize automatic inspection and patrol in the vault. The robot is equipped with a variety of sensors and cameras that can monitor the environment inside the vault in real time and complete physical inspections efficiently. When an abnormal situation occurs, the system can automatically dispatch emergency response, such as automatically closing the vault door, etc., to improve the ability to respond to emergencies. At the same time, the robot can also interact and authenticate with employees in the vault through voice interaction and face recognition and other technologies, and inspect whether the staff has illegal operations in the vault.
RFID tag technology is used. Using IoT technology and RFID sensors used on the chest, the location and status of cash, precious metals and other items in the vault can be monitored in real time. Inspectors can quickly and accurately take stock of cash and precious metals stored in the vault, avoiding the tedious and time-consuming nature of traditional manual counts. At the same time, the unique identification code of the RFID sensor can be used to trace the circulation history of the item, which is convenient for finding and checking.
Luggage detection and early warning technology. Set up a special cross-beam high-definition camera near the safety protection door, multi-angle monitoring of the entry and exit personnel, identify and detect whether to carry bags, once suspicious bags are found, they will be reminded by the sound and light alarm equipment on the scene and notify the monitoring center on duty in real time.
Automatic detection of anomalies in the library. Using video intelligent analysis technology, all monitoring points in the warehouse are deployed to realize real-time monitoring and early warning of violations such as personnel entering the warehouse without work clothes, less than 3 people in the warehouse, abnormal lighting in the warehouse, and unaudited personnel. After the system detects abnormal events, it will capture them in time and push relevant records to the management personnel.
Surveillance video concentrating playback and spot check technology. AI technology can automatically generate a complete video recording of a person's trajectory in one multi-channel video. That is, with a person's behavior trajectory as the main line, through the main screen + small screen, multi-angle monitoring is realized, and multi-channel surveillance video is integrated on one screen to achieve 360-degree behavior display, which greatly improves the integrity and timeliness of monitoring and access, improves the convenience of inspection, and solves the tedious work of manual switching of multiple monitoring screens for traditional inspection. In addition, it can realize the automatic compression of the still picture of the surveillance video, which effectively solves the work efficiency problem of video inspection and search.
The challenge of intelligent security technology in the current large-scale promotion of vault security management
First, the level of technological maturity is not high enough. When new technologies are first applied, they may have problems such as unstable performance and imperfect functions, which need to be gradually improved and improved after a certain amount of time and practice. In addition, the constraints of R&D manufacturers on factors such as vault application scenarios and algorithm requirements, as well as the comprehensive consideration of financial units in terms of investment and output, test time, etc., make it difficult for intelligent security technology to achieve general standards suitable for various business libraries and promote them on a large scale.
Second, there are certain security and privacy risks. With the advancement of technology, treasury risk prevention needs to deal with more and more cyber and information security risks. Issues such as cyber attacks, data breaches, and system failures can lead to certain threats to vault security.
The third is human factors and legal and regulatory risks. The introduction of new technologies requires training and education of treasury staff to familiarize them with the operation and application of the new system, as well as overcoming issues of acceptance and adaptability of personnel to new technologies. Intelligent security of vaults involves the core assets and customer information of financial institutions, which need to comply with relevant laws, regulations and regulatory requirements to ensure compliance and legality.
To sum up, the application of intelligent security technology in the treasury management can timely identify the abnormal behaviors and abnormal situations brought by personnel, the on-site environment and equipment, which can better improve the ability of the vault to prevent and respond to various risks and hidden dangers, and will make the intelligent security management of the vault move towards a new stage of more efficient, intelligent and safe.
(This article was published in the second half of March 2024)