What is the use of video analysis.
Video analysis involves the use of software algorithms for analyzing CCTV images in order to determine the factors when an alarm signal needs to be sent.
Registration of the time mode in such video surveillance systems can be automated.
Video analytics is implemented in special software that is capable of selectively recording objects with specified parameters.
The technology helps to track objects taking into account the perspective set by the user.
The probability of false alarms is the lowest compared to other video surveillance systems.
Video analytics is an analysis of video information, as a result of which objects of interest for performing certain tasks are recognized in the video image. This allows you to assign certain «behavior models» to the recognized objects.
If the model is violated, the system sends an alarm signal to the operator, who then decides on further actions.
Video analysis is usually used to solve the following problems:
1) tracking an object – it is possible to track an object located in the observation zone of a camera, which is capable of automatically changing its direction and, if necessary, enlarging the image of the object for detailed identification;
2) Determination of the level of crowding: if the permissible limits of crowding are exceeded in certain areas of the camera's field of view, the system issues an alarm message to the operator;
3) Search for video by face image in the archive: the system carries out a quick search in the video archive by the face image saved in the video archive. The signal is sent when the desired face is detected in the camera's field of view in real time;
4) Counting visitors: the number of people entering and leaving;
5) Registration and recognition of car license plates.
Motion Detector Program
One form of video analytics is the motion detector program.
This program allows you to take pictures from a WEB camera if movement occurs in its field of view.
The pictures are saved in the specified location on the computer, then the picture itself and a report on the movement that occurred are sent to the mailbox.
Motion Detector Programcan also work in hidden mode: the program is not displayed on the screen.
There is a minimum list of requirements that a modern motion detector program must meet: multi-zone with the ability to select specific zones.
To implement this function, it is enough to learn how to work with a video archive, selecting an event in a specific time period.
A higher level of analytics is the ability to find not what is prepared in advance by the event log, but to create a new event based on the existing saved information.
For a quick search in the archive, it will be necessary to save the information flow from the motion detectors in advance, and not only the information that triggers an alarm signal according to the configured parameters, but also any activity.
Such information can be recorded in a different resolution, determining the accuracy of the detector, in the database.
Later, when searching, we select any area of the camera's field of view and set up a search in the database using the detector data, obtaining video recordings that interest us — as opposed to events prepared by the log.
The motion detector program at a high level implies preliminary data processing by «intelligent» motion detectors. They increase the possibilities of working with video archive data.
Such detectors do not work with activity in specified zones, but with objects that have certain characteristics, which they calculate.
At a minimum, such characteristics will include: movement vector (direction and short-term speed), dimensions, history (path, or track of the object), less often — the type of object (person or group of people, vehicle).
Alarms are set already by violations of the objects of the nature of their movement in specified areas, i.e. the concept of a zone is also acceptable here, the concept of a «border» is also often encountered — a virtual line, the intersection of which in one direction or another (and the characteristics of the object) will cause an alarm or can simply count objects.
Outdoor video surveillance online
Simple detection is used only indoors.
A complex form of video analysis is used mainly in outdoor online surveillance systems.
Outdoor video surveillance onlinewith the help of this program is carried out with a much lower percentage of false positives.
The work of video analytics in this direction consists, first of all, in working with objects.
The archive is searched for movement with a set of specified properties in the frame.
The area monitored by the video camera can be divided into several sectors by interests.
Online street video surveillanceallows you to call events in accordance with one type of settings, and the video recorder records the parameters of all movements and objects in the database.
Online street video surveillance in the video analytics system is usually used for perimeter protection.
Any boundaries are covered: water, land, alienation zones, etc.
The viewing area of the video camera guarding the perimeter can be designated by boundary lines.
If an outsider, whose parameters are pre-set, crosses these boundaries, the system gives an alarm signal.
Online street video surveillance in the video analysis system allows you to create «exclusion zones».
These are areas in front of the protected boundaries, upon the appearance of a subject on which there will be signal has been given.
Subject information can also be archived and used for searching.