Application case of intelligent video surveillance system at San Francisco Airport, USA
chrispeng [ad_1]
System design requirements
San Francisco International Airport is one of Vidient’s earliest customers. After “9.11”, the United States established the Department of Homeland Security, which is unified in charge of the security of its homeland. As one of the largest international airports in the United States, San Francisco International Airport has also vigorously strengthened security measures. San Francisco International Airport has installed more than 1,600 cameras, covering almost every corner of the airport. However, due to the limited number of security personnel, each person has to monitor dozens or even hundreds of cameras on average. In addition, because there is no intelligent recognition function, most of the videos are only stored for later query, and real-time warning cannot be achieved. In early 2005, a security incident occurred at San Francisco International Airport. A passenger broke into the terminal without passing the security check. The airport had to close the entire terminal building. Although there are video recordings, it took more than half an hour for airport security personnel to find the passenger from the recordings and catch him because there is no intelligent index. In this incident, the airport lost more than one million U.S. dollars.
In order to further strengthen intelligent security monitoring, San Francisco International Airport in the United States adopted Vidient’s SmartCatch intelligent video surveillance system solution. So far, the installation, commissioning and acceptance of 18 cameras in the first phase of the project have been completely completed. 500 cameras will be installed in the second phase of the project. After using SmartCatch, the airport discovered that there are hundreds of incidents threatening security every day, which have not been discovered before.
System design, function, structure
The 18 monitoring points of the technical design of the first phase of the project are as follows:
1. Classification of camera functions
●Detection of orientation movement at airport exit: 2 cameras;
●Illegal trailing detection of access control system: 7 cameras;
●Unattended object detection: 5 cameras;
●Conveyor belt baggage jam detection: 4 cameras.
2. System equipment
●Camera, network: The airport has already installed the camera and network, SmartCatch is fully compatible with them, no longer need to install;
●Video capture cards: 18, because their cameras are analog signals, they need video cards to convert into digital signals;
●PC: 7 units, used for video processors (VPUs), theoretically speaking, a PC can drive 4 cameras. But in order to ensure the speed, two more PCs are used;
●Server: 1 set.
3. Software
●SmartCatch VPUs: 18;
●SmartCatch server: 1 pc.
Intelligent video surveillance system can realize behavior recognition
Why do we need intelligent monitoring?
●Too many screens: There are too many screens to watch, and more than 98% of surveillance videos actually don’t need to be paid attention to;
●Human limitations: human factors, other tasks;
●Impossible tasks: Many surveillance systems have 100 or 1000 cameras and cannot observe everything that is happening;
●Overloaded video volume: There are too many data to be paid attention to at each moment, and it takes a lot of manpower to find past records.
1. Recognizable video behavior
Crowd density detection: detect the crowd density and the length of the line to prevent overcrowding;
Oriented motion detection: it can detect people, cars, and various objects in violation of the rules;
Person trailing detection: detect and track the illegal trailing of people at the entrance guard;
Loitering detection: detect whether someone is loitering in a sensitive area;
Peripheral intrusion detection: stable detection, tracking and classification of invading people, vehicles or objects;
Removed object detection: detect whether the object is removed;
Parked vehicle detection: detect the size and time of vehicles parked in sensitive areas;
Turnstiles violation detection: to detect whether someone jumps or goes through the turnstiles;
Unattended object detection: detection of discarded and unattended objects (suspected explosives, etc.);
Vehicle trailing detection: detect and track the illegal trailing of people or vehicles at the entrance of the vehicle.
2.System Features
Fully automatic real-time video monitoring: real-time monitoring, analysis and interpretation of the video, when specific security threats are detected, the system will automatically alarm the corresponding security personnel.
Accurately monitor security threats: SmartCatch can accurately detect scheduled violations of security regulations. It can work in a variety of complex actual environments such as indoors and outdoors, stably achieving the high recognition accuracy rate in the current industry.
Policy-based response management system: SmartCatch can be easily configured according to the user’s policy requirements, and take effective response and upgrade measures to alarms.
Compatible with existing monitoring systems: SmartCatch is compatible with all CCTV systems on the market, and it is easy to integrate with other sensors and security management platforms.
Easy to install and use: SmartCatch can be easily configured in accordance with user security policy requirements, including when, where and how to send alarm information, alarm upgrade notifications, and send alarm information through multiple channels, etc. SmartCatch can monitor multiple video inputs at the same time, display alarm details, and provide a list and statistics of all previous alarms.
3.System performance
Accuracy: SmartCatch can consistently achieve 95% to 98% accuracy in actual scenarios. By using the most advanced machine learning, multi-object tracking, and behavioral reasoning technology, SmartCatch can identify a variety of behaviors that threaten security.
Multiple behavior recognition: SmartCatch can recognize more than ten kinds of behaviors, which can meet the different needs of indoor and outdoor, and provide comprehensive and complete solutions for the entire enterprise.
Comprehensive response management system: SmartCatch’s alarm response management system can provide security personnel with real-time, fast and accurate response management strategies, and can be adjusted according to the security system.
The versatility of the solution: SmartCatch can be used on several cameras, or on thousands of cameras. This can be designed according to user needs. SmartCatch is compatible with existing CCTV camera systems, and is easy to integrate with other sensors and security systems to provide users with global monitoring.
4.System functions
Video processing capabilities: multiple object recognition, multiple object tracking, day/night self-regulation, automatic filtering of environmental noise, automatic filtering of background noise;
Video panel: Real-time video playback can be performed on detected objects or events, and multiple cameras and alarm events can be observed at the same time;
System scalability: the system can be used for an unlimited number of cameras, each camera can be tested for multiple behaviors, and an unlimited number of users can use the system;
Supported input sensors: access control signals can be input through Dry Alarm Contact, and other sensor signals can be input through application programming interfaces (APIs);
(PTZ) Camera control: You can directly guide the PTZ camera to zoom in and observe suspicious behavior, and you can only focus on the user’s predetermined area;
Supported alarm output: SmartCatch window output console, telephone, mobile phone, handheld computer, pager, email and Internet, sound and light alarm, any other alarm device that supports API;
Alarm processing management: Alarm upgrade can be configured according to time, personnel to be notified and priority, each alarm can be sent to multiple security personnel, real-time video playback can be performed, and the alarm location can be easily displayed;
User control platform functions: visualization of event status, simultaneous display of multiple alarms, display of each alarm’s physical location and type, real-time display of the status of each alarm, manual change and setting of alarm status and upgrade strategies;
Administrator functions: set users and manage user groups, set alarm upgrade strategies, set cameras and alarm schedules, and easily change strategies, users, user groups, and schedules;
Report items: alarm summary report, alarm accuracy report, missing alarm report, all event report.
5. System platform
The SmartCatch intelligent video surveillance platform provides a unique distributed architecture, which can be used for several cameras or thousands of cameras for the entire enterprise. It provides more than ten kinds of behavior recognition, which can be used alone or combined together to provide a comprehensive solution for the enterprise. SmartCatch provides a complete set of application program interfaces, which can be easily integrated, customized, and other programmable applications. It is compatible with existing CCTV systems and security facilities, and can be used in conjunction with access control systems, biometric identification, RFID, etc. Alarms can be sent to various devices in real time, such as mobile phones, pagers, e-mails, PDAs, etc., or they can be logged in and viewed from any corner of the world via the Internet. SmartCatch runs on industry standard hardware and operating systems.
6.System Requirements
SmartCatch server
●Operating system: Windows 2000/XP or Linux;
●Processor: Intel Pentium 4 3.2 GHz or Higher;
●Memory: 1 GB RAM;
● Hard disk: 120 GB.
SmartCatch Video Processors (VPUs)
●Operating system: Windows 2000/XP or Linux;
●Processor: Intel Pentium Xeon 3.02 GHz or Higher;
●Memory: 2 GB RAM;
● Hard disk: 80 GB;
●Maximum video input: 4 / VPU;
●Frame frequency: 30/25 fps NTSC/PAL;
●Camera type: digital/analog/network;
●Maximum image resolution: 640*480, 60Hz 4-channel PCI image compression card.
[ad_2]
