Video analytics are not all the same.
The reason for the error is ignorance of the best.
Democritus
When deciding to implement video surveillance systems, company executives often believe that the technology is capable of solving all the tasks of protecting an object. Is this true? Of course not. For example, one of the main problems of many of the existing video surveillance systems remains the need to work with almost each camera of an individual operator. Otherwise, the video surveillance system becomes either meaningless or serves only to archive information. The operator's work, of course, must be paid, and this significantly increases the cost of owning the system.
What do we expect from modern video surveillance systems? They should be informative, and the information should be sufficient for human decision-making. To eliminate the influence of the human factor, security system developers offer video surveillance systems with video analytics. The tasks of such systems are to free the operator from analyzing the video stream, offering him only the information that is necessary for decision-making. A person is a bad observer, but he makes decisions perfectly.
The rapid increase in demand for intelligent video surveillance systems has led to the fact that currently almost all players in the security systems market offer video surveillance systems with video analytics. However, the term «video analytics» is often understood to mean things that are far from each other. There is confusion in the security systems market in terms that, in my opinion, should be sorted out.
I hope that this article will help you understand what is called video analytics today and what systems are needed by consumers when they talk about the desire to install a video analytics system at a facility.
Video analytics – motion detection?
A massive transition to digital solutions is taking place in all industries related to the processing and transmission of information. Quite recently, digital systems could not compete with analog systems, and information was recorded and transmitted only in analog form, but progress in electronics is leading to a transition from analog to digital. In the field of security systems, there is also a transition from analog video surveillance to IP systems. Open standards are used in IP video surveillance equipment, which simplifies the installation and modernization of equipment, and it becomes possible to use products from different manufacturers in one video surveillance system.
One of the main advantages of IP video cameras is the presence of built-in motion detectors. A motion detector is a module whose main task is to detect the movement of objects in the camera's field of view. Motion detection is a simple form of video analytics that has a number of significant limitations. For example, sudden changes in lighting conditions, precipitation, camera shaking, the appearance of birds and insects in the camera's field of view lead to false alarms. For these reasons, systems with motion detection are usually used indoors.
Most video surveillance system providers refer to motion detectors built into cameras as video analytics.
As a rule, manufacturers of video cameras with motion detection offer free software (SW) for managing the system. However, it is worth understanding that free software is intended mainly for monitoring objects. Recording in most free programs is organized quite primitively. If free software allows you to work with several cameras and has functionality suitable for home use, then paid software for video surveillance systems provides advanced features, such as managing any number of cameras, integration with other security systems.
Video analytics — computer vision?
Some manufacturers understand video analytics to mean much more than motion detection via IP video cameras. Currently, the security systems market offers video surveillance systems with computer vision. In addition to monitoring and archiving events, such systems recognize dangerous situations in real time and automatically provide information about them to the operator. All he has to do is make a decision. Due to this, the security service of even a very large facility equipped with hundreds of video cameras does not require a large staff of operators.
These systems eliminate the influence of the human factor and operate on the principle of “one object – one operator.”
Video surveillance systems with computer vision recognize objects (people, cars) and situations such as fire, smoke, abandoned objects, thrown objects in real time from still video cameras and automatically point rotating video cameras at them to obtain a detailed image. All events are displayed on the object map. The video analytical map and detailed video image can be displayed on a monitor, video wall, terminals supporting Touch Screen technology, and on a PDA.
The archive search is performed by time, target class, camera number, event, operator comment. Video surveillance systems with computer vision can use analog and IP cameras, as well as thermal imagers, which allows upgrading previously installed video systems at the facility and switching to new video analytics capabilities.
To digitize the video stream from analog cameras, it is possible to use video capture cards installed in the recognition server, or additional equipment — encoders that digitize the analog signal of the video camera into a digital stream, which is transmitted to the recognition server.
Computer vision video surveillance systems are capable of working together with perimeter security sensors, automatically providing a video image from the area where the sensor was triggered. In this case, the operator's screen displays an image of the corresponding fixed camera, the rotating camera is automatically aimed, and target tracking mode is possible. The video analytics map displays information about the occurrence of this event.
An important feature of video surveillance systems with computer vision is not only the recognition of objects and alarming situations, but also the ability to automatically tune out false alarms when cloud shadows appear, bushes and trees sway, insects appear in the camera field, video cameras swing and shake, and car headlights illuminate the area. Video surveillance systems with computer vision are able to adapt to external conditions without operator intervention.
The use of such video analytics in systems allows in many cases to ensure the protection of the perimeter and territory of large facilities.
Market trend – absolute optimum
The process of searching for optimal solutions in the field of video surveillance systems is almost continuous. And it brings results.
Domestic and foreign developers offer video surveillance systems built on a modular principle. Their basic functionality can be optionally supplemented with intelligent modules that allow for recognition of vehicle license plates, manual or automatic control of PTZ cameras, tracking of detected targets by PTZ cameras, object tracking, face recognition, notification of an object appearing in a specified area, detection of thrown objects, etc. Such video surveillance systems are able to best meet the needs of facilities of various scales and industries.
That is, today universal IP security video surveillance and automated business monitoring systems have appeared on the TSB market.
These systems implement a wide range of video surveillance capabilities: from a simple video recorder to advanced video analytics.
What are the advantages of such solutions? Firstly, it is now possible to equip an object with any video surveillance equipment using a single universal platform. Secondly, such systems allow you to create the required number of video channels and functionality of the video surveillance system for a specific object or task. For example, a channel with video recording functions, a channel with video analytics functions (motion detection), a channel with advanced video analytics functions (computer vision), etc.
The user chooses what he needs, without overpaying for «extra» video channels and unclaimed functionality. By choosing channels with video recording functions, the user receives a fully functional video surveillance system with the ability to view «live» video, record and play back archives, connect PTZ cameras and control them manually. Channels with video analytics functions (motion detection) allow you to create a video surveillance system with the ability to detect moving objects, display their trajectory and transmit information about penetration into the protected area to the operator in automatic mode. Video channels with advanced video analytics allow you to create a video surveillance system with computer vision, with the ability to automatically detect and recognize targets (people and cars) and alarm situations with the transmission of video information about emergency situations to the operator in real time.
Such video surveillance systems can be integrated with license plate recognition systems, perimeter security radar stations, video detectors of thrown objects, access control and management systems, and any sensors. This allows you to get a universal platform for comprehensive security in one system.