complex security system
To date, the company has installed about half of the total number of active video cameras in the largest regions of the country. Currently, there are about 45,000 access cameras, 9,000 courtyard cameras, and 500 cameras in crowded areas. Access is provided not only to law enforcement agencies and city structures, but also to ordinary residents. The participation of citizens in the project will give a qualitatively new result of using video surveillance systems. There is also a transition from simple recording and recording of events to more active and intelligent algorithms of the city’s video surveillance system.
Smart City system
The Smart City project is a hardware and software complex using automatic video data analysis methods, which online helps to identify and prevent violations of law and order on the roads and yards, helps to identify fires and fires at an early stage, controls lighting and cleaning streets and much more.
Goals and objectives of the implemented program:
• ensuring the protection of public order and safety;
• increasing the efficiency of interaction and coordination between the services of public law and order;
• suppression of facts of violation of public order;
• the ability to quickly receive information from urban infrastructure facilities and access to data storage servers.
Typical objects in the Smart City data control system:
• residential sector (entrances, courtyards);
• municipal buildings, schools and kindergartens;
• industrial facilities and energy complexes;
• objects of transport infrastructure (roadways, stops and saloons of public transport, train stations and airports, subway);
• places of mass gathering of people.
description of the scheme of work
We can distinguish two prevailing most applicable architectures – centralized and decentralized. In the case of building a system according to a centralized scheme, all data from sensors flows into a single center for collecting and processing information. Other infrastructure services and networks also have access to a single data center. Using a decentralized architecture, data from sensors flows to the data collection and processing center. Information is requested and the collected data is accessed by third-party infrastructure networks using specially designed application centers.
The city video surveillance system includes three most popular data processing algorithms:
1. A system of service sensors for tracking and monitoring the performance of cameras. The sensors detect such malfunctions as signal disappearance, failure of the automatic image adjustment system, and contamination of the zoom lens. Service sensors automatically detect abnormal situations, for example, camera obstruction or lens closure, violation of focusing systems, camera orientation change and deliberate illumination of the matrix.
2. The system of indexing sensors to optimize the search for the data archive. The presence of metadata in the event storage database significantly increases the efficiency of the work of the circulating infrastructure authorities when establishing the relationship between events. Using tags, an employee quickly finds the necessary frames of the video archive, which is a hundred times more efficient than manual search. The ability to search “by event tags” greatly simplifies the work of various organizations when analyzing data, and system operators when identifying and classifying events.
3. Behavioral analysis of emergency prevention. Analysis of data received from cameras installed in public places allows automatic recognition and signaling of such alarming situations as:
• violation of stopping places of vehicles;
• crowding of a large number of people;
• accelerated movement of people and objects;
• violation of the operating mode of objects restricted to access;
• falling of people on the rails;
• forgotten and intentionally abandoned things and more.
Data processing board “Accelerator ST325”
The processing and analysis of data is based on a large list of algorithms implemented on a specially designed and manufactured board “Accelerator ST325”.
Interfaces: PCIe x1 v. 2.0, Fast Ethernet, SERDES, UART. USB
• Memory: SDRAM 256Mx64
• Programmable generator from 12 to 300 MHz
• Power consumption: 40 W
• FPGA Xilinx Kintex 7 XC7K325T
Various types of memory modules are installed on the board:
• SDRAM up to 128 MB;
• FLASH NAND to support the LINUX OS;
• EEPROM ensuring OS boot.