In certain situations, camera surveillance systems require better video analytics to meet specific monitoring and recording needs.
In this article, you will learn what are the advanced video analytics capabilities that are available in the market and how they can be used in day to day situations.
==> I also recommend reading my other article on the use of basic video analytics technologies such as motion detection, and virtual line. Click to read the article: Intelligent Video Analytics Cameras and Software.
Advanced Video Analytics
Assuming you already have the basic knowledge about video analytics and you've probably read the other article on the subject, let's get right to the point and talk about the most advanced features. The following is a description of each of them.
Face detection and recognition
This type of resource is very common in places where it is desired to control the access of people to certain restricted areas or as a way to later identify who had access to a specific place.
Security cameras are installed at specific entrances or places where there are movements of people who must have a picture of their faces inserted into a database of a security system.
Although there are complex algorithms behind the software responsible for recognizing faces, the practical application concept is quite simple, a security camera must identify what a face is and capture an image as clearly as possible to store it in a database for later consultation and comparison with people registered in the system.
There are cameras or software on the market that are able to perform only the detection of faces and others that can, in addition, capturing the image of faces, make the comparison with a database to find and identify people.
Vehicle license plates recognition
This type of advanced video analytics is very common on roads, tolls, and areas where there are vehicle entrances. The process of recognizing vehicle license plates numbers can be done on the camera itself or in a combination of camera and specialized software.
A few years ago it was possible to recognize vehicle license plates only through specialized and expensive software developed by a few companies that dominated the electronic surveillance industry.
Over time and with the advancement of technology more options have appeared in the market and the price of implementation of such systems has been reduced, besides the software, there are also IP cameras that do such a job themselves.
Companies such as ISS and Genetec stand out in the market for having robust and accurate systems for vehicle license plates recognition, but they are not exclusive because with the opening of the market there were many partnerships and commercial agreements that allowed the use of the intelligent analytics algorithm by other companies that implemented this functionality into their own recording and monitoring software.
Currently, it is possible to buy security cameras that have embedded intelligence to recognize vehicle license plates and the price is very affordable. The Chinese manufacturer Hikvision has IP camera models that were designed for this purpose.
In order to capture vehicle plates numbers, however, it is necessary to follow specific camera installation and positioning procedures with adequate angles to obtain images of sufficient quality for analysis and recognition by the camera itself or by a server that runs a specific software.
Tracking the movement of people and vehicles is also something that can be very useful as it allows you to analyze behavior and record occurrences automatically based on specific, pre-established events.
There is intelligent software that can accurately track the movements of people and vehicles and generate alerts for an action to be taken immediately or to record the event with date and time and allow smart searches to be performed later.
Here is an example of the practical application of this concept.
Note in the example of the image that a car is positioned in the pedestrian crossing area and there is an indication of tracking a person walking through it, this type of intelligence allows to alert and record incidents with great accuracy without the need of a camera operator looking to the monitor.
Another practical example of using this video analytics technology is the control of vehicle movement at a particular intersection where it is allowed to travel in specific directions.
See the following image where a vehicle converts to the left and is being monitored and tracked by an intelligent system, in which case it is possible to count the number of conversions in that direction.
Tracking technology is very useful in situations where there is an interaction between people and vehicles and alerts can be generated with great accuracy.
A very common application in intelligent monitoring systems is the counting of vehicles or objects that enter a specific area or cross a virtual line.
Counting vehicles at entrances and exits
The concept of video analytics used here is very simple, a virtual line is drawn on the camera image to count the entrance of cars in a certain direction and the same process is performed for the vehicles that are leaving the place, changing only the direction in which the vehicle moves.
Note in the image that the vehicle is crossing a virtual line and there is an indication of the count. A second virtual line is positioned at the entrance for the same job.
Counting vehicles on roads
It is possible to use advanced video analytics to perform counting of vehicles on roads and highways and record the direction of traffic. See the following image for a practical example of this application.
Note that there are virtual lines that record the passage of vehicles in both directions, some caution to be taken in this case is with the control of the environment because the conditions of the lights are variables that can seriously affect the accuracy of the counting process.
Imagine a situation at night with rain or fog where you can not clearly identify the vehicles that travel through the road or the use of headlights end up affecting the final result of the analysis.
Some specific cases can be identified as behavioral analyzes, a person approaching a fence or wall to try to jump to the other side or lowering himself to perform some suspicious activity such as digging or putting something on the spot.
The combination of advanced video analytics such as tracking, crossing lines and entering and exiting areas allows the generation of events that can be classified as behavioral analysis.
In some cases, this type of video analysis is already part of the software and in other cases, it can be customized as the combination of different functions.
See the following image for intelligent video analytics that combines the use of tracking of a person approaching a fence and entering a specific area and a possible crossing detection line if the subject tries to jump over the fence.
Through intelligent video analytics, it is possible to identify people or objects that have been standing for a long time in a specific area, this function is known as loitering and can be configured in IP cameras and software that can identify this behavior.
Note in the following image that an object was abandoned in a demarcated area, the system simply identifies that there has been a change in the scenario and after a certain time it classifies this situation as loitering, generating an alert for the operator or sending messages to a particular person or group.
This feature is very useful for identifying potential suspects close to restricted areas or even identifying that an object has been abandoned at that location.
Abandoned or removed object
If an object is left in a previously demarcated area the camera or monitoring system can identify this action as abandoned object, which in practice turns out to be something similar to the previously explained loitering function, however with the difference that if the object does not move for a long time it passes into a "static" state (idle) that characterizes abandonment.
When you remove this static object from a certain area, the video analytics system can identify the action as "object removed". See the following image with an example of this application.
In this example, the person drags the object that was placed in the demarcated area and triggers the removed object event.
The names of the types of intelligent video analytics may vary depending on the software developer or security camera manufacturer, in many cases, the function is called virtual fence. Notice in the following image that a person tries to throw an object over a fence and an alarm is generated automatically.
The event is generated by crossing the object by a virtual line, so it is possible to find the same function with the name of virtual line crossing in different software or cameras.
Identification of containers
A system that uses advanced video analytics with optical character recognition (OCR) can read texts written on the side or top of containers and thus identify them automatically.
This intelligent video analytics is used by transportation companies from different parts of the world and can be implemented through the use of specialized software such as SecurOS Cargo from ISS.
Imagine a situation where there are a lot of people in a specific location, with the intelligent video analytics system it is possible to identify a crowd.
This function is part of the ISS SecurOS Tracking kit and although it is similar to what was previously described as loitering, there is a certain difference because the software does a kind of behavioral analysis.
There are numerous types of advanced video analytics that can help identify specific events that are classified and named by different companies that develop or adapt the algorithm for particular applications.
Obviously, you will find other types of video analysis described otherwise or with other names, but the important thing is to know how to assess your real need in the application of such technologies.
I hope this article can help you make a better decision when choosing the right software for your application.
Want to learn more ?
If you want to become a professional CCTV installer or designer, take a look at the material available in the blog. Just click the links below:
Please share this information with your friends...
Claudemir Martins is a former Samsung Engineer with 16+ experience in the surveillance industry. He has been traveling around 17 different countries to teach people how to design, and install CCTV systems.