4 billion rubles for an intelligent video surveillance system for the Moscow Metro

At the end of last year, a competition was held for the development of project documentation for equipping the Moscow Metro facilities with an intelligent video surveillance system (ISVN). The amount of the contract amounted to more than 387 million rubles. And this is only 10% of the total amount of approximately 4 billion rubles allocated from the federal and city budgets for the implementation of the most advanced intelligent video surveillance system in the Moscow metro. Large sums and far-reaching plans, in fact, are the resumption of the state program, which was in force from 2010 to 2013 and started in connection with all of us known tragic events that occurred in the Domodedovo and Moscow Metro. Let's see what has been done over the years, and what we should expect from the security program in 2014-2016.

Ambition and Innovation

Perhaps, it’s worth starting with the fact that in Russia safety in the metro, as in many other countries, is under the jurisdiction of the police. On the shoulders of ordinary employees - exclusively operational issues. Nevertheless, we know that the police are now going through optimization (read: reduction), and the number of passengers in the metro of the capital is unlikely to decrease.



In 1998, the Moscow Metro acquired the first video surveillance systems with video recording. If you think positively, it is now difficult to calculate the number of administrative and criminal cases in which these records appeared. Today cameras are installed at all stations and in all crossings, as well as in all cars of the Koltsevaya line.



On the other hand, a huge amount of video remains unvisited due to the banal human factor: none of us can monitor several screens at the same time, and even several hours in a row.

')

Therefore, the idea of ​​implementing the most ambitious video analytics project in Russia and the CIS, and from the budget point of view, and from the technology point of view, the introduction of intelligent video analytics in the field of transport, has emerged.



Since the middle of 2012, at many stations of the Moscow Metro, pilot deployments of intelligent video surveillance of both domestic and foreign vendors have been carried out. And so, on the basis of the data obtained, the metro prepared an impressive technical task for the development of design documentation for equipping WIWN 188 stations, 15 depot, tunnels, electrical substations, traffic control points. The Terms of Reference provides several levels of security for various categories of metro facilities. In addition, now, in addition to the internal metro services, the external users from the departments of the FSB of Russia, the Ministry of Internal Affairs of Russia, the Ministry of Transport of the Russian Federation, the Emergencies Ministry of Russia, will be connected to the system of intellectual video surveillance (as well 24) the system will be connected to the Moscow City Control Industrial Administration. In general, everyone will ensure our safety, except, perhaps, the customs authorities.



The system of intellectual video surveillance, unique for our country, will unite in itself six main subsystems, the tasks of which you can find in the table:

Subsystem	Tasks
Survey and situational video surveillance	Automatic detection and reporting of the following events: falling of people and objects on traveling space unauthorized passage of passengers to the forbidden areas of the metro (tunnels, service premises, etc.) crowding movement in the forbidden direction, including against the main stream of people fast motion (running) appearance of abandoned items
Biometric CCTV	Automatic detection and recognition of passengers' faces in areas of biometric video surveillance, including: automatic capture of images of all passengers automatic detection of the facts of passage through the turnstiles without payment and automatic fixation of photographs of "free riders" recognition of persons passing through the line of control, automatic identification of persons included in the list of violators and other lists generation of an alarm signal about the fact of fixing a person on the list of violators and other lists prediction based on the time of fixation of the recognized person in the areas of biometric video surveillance of the trajectory of movement of the fixed face
Informational - analytical subsystem	Providing functions of analyzing statistics of passenger traffic (in the form of tables and graphs): total passenger count counting passengers without fare queue length time in line number of people in the zone (crowds)
Management and monitoring	The subsystem should provide: equipment control detailed information on each monitored device monitoring the quality of the video streams provided localization of failures and malfunctions of devices; notifying service personnel and providing necessary information for prompt troubleshooting
Patrol / Outfit Management	prompt notification of all alarm events in the patrol \ outfit area Operational verification of the database of persons violators, including photographing a suspicious person, the formation of a request to the subsystem of biometric video surveillance and automatic verification of lists of previously recorded offenders
Information Security	Identify the main types of information security threats: illegal actions of third parties erroneous actions of users and attendants software failures and crashes malicious software and hardware effects on computing equipment and information

TZ, it would seem, is not easy: today, no company in the world is able to solve such a huge list of tasks on its own. That is why the video surveillance system of the metro will consist of solutions from different suppliers, but still represent a whole, in which all subsystems interact with each other.



The issue with the integration of equipment and software is solved quite simply: for the first time in the history of government projects of this magnitude, ONVIF is stated in the technical task as a mandatory requirement that complies with the law on the transition of federal institutions to the use of free software. It is intended to use ONVIF for broadcasting live and archived video, metadata and video analytics events, as well as for centralized management of all components.

Data is everything.

Comparison of information received from surveillance video, biometric and situational video analytics, will allow you to automatically detect illegal actions, fix the offenders and check them on the bases of the search. And also to predict the route of their movement for the implementation of the detention by the security service.



To perform these tasks, two types of cameras will be used: for surveillance and situational surveillance - IP cameras, and for biometric surveillance - machine vision cameras.



The information-analytical subsystem also plays an important role, because the metropolitan metro passes more than 8 million passengers daily. Obtaining statistical data on the distribution of passenger traffic will allow you to adjust the work of escalators, the intervals between trains, the work of cash desks and attendants.

For sure!

In accordance with the “Object categorization system according to the degree of potential danger and terrorist vulnerability”, the Moscow Metro State Unitary Enterprise falls into the First category and imposes strict requirements on the accuracy of video analytics at the site. The requirements are:



Requirements for the accuracy indicators of situational video analytics

Characteristic	Value
probability of true positive positives	not less than 98%
probability of false positives	no more than 0.5%

Requirements for accuracy indicators of biometric video analytics

Characteristic	Value
probability of face detection in passenger traffic	not less than 95%
probability of truly positive person identification	not less than 85%
probability of false positive identification	no more than 0.5%

The terms of reference do not describe the methodology for testing and evaluating actual indicators of video analytics accuracy. Most likely, here the TZ developers relied on the methodology of the most extensive tests of video analytics conducted at Novokosino station.

3D and priority

To justify the location of the IPS cameras, assess the accuracy of video analytics and eliminate dead zones, it is proposed to carry out three-dimensional modeling of metro objects and camera coverage areas.







The Moscow metro is so crowded that the inevitable daily detection by the video analytics of a huge number of events. Some of them do not require a lightning reaction from the staff of the metropolitan and external services, but there is another category of events, for example, the fall of a person on rails or the discovery of a dangerous criminal. In order to respond to the most important, impatient events, events in the first place, the system of priorities ( event ranking ) will be provided in the system of intellectual video surveillance . The priority of an event can depend on various signs: the type of event, the observation area, the accuracy of recognition, the distance of the object and others, as well as a combination of these factors. And for different categories of users, priorities will be different, since each service has its own area of ​​responsibility.

Mobilization of "mobile phones"

In order to promptly receive information on alarm events, patrol services that monitor order in the metro area, as well as other metro officers, will be equipped with mobile terminals. They will receive prompt notifications of alarming events that require quick response, including the appearance of persons of interest to the patrol. When such a person is detected, the intelligent video surveillance system will determine the possible trajectory of its following and send a notification to the mobile terminals of the orders located in the immediate vicinity of this trajectory. The notification will display a photo of the recorded person, the time and place of its detection, as well as the reason for attracting the attention of the patrol and instructions for actions. The delivery time of an alarm event to a mobile device will be no more than 10 seconds, and information will be transmitted using wireless Wi-Fi access points and using GSM 2G / 3G networks.

Metro 2033

Of course, the Moscow Metro still has to do a lot of work on preparing regulations and personnel training, and it is not clear how the metro infrastructure will be used, because all the documentation on the intelligent video surveillance system is focused on technology. But there is no doubt that the ISVN will significantly increase the safety and quality of the metro.



Author: Nikolay Konokhov

Director of Development for Security in Synesis