Intelligent video analytics is very important in modern CCTV systems because it allows the identification of real-time occurrences and the rapid search for recorded events, which facilitates the operator's life and allows a faster and more dynamic response to a possible problem.
In this article, you will learn what intelligent video analytics is, what are the most common types, and how to use it in specific situations that require faster event identification.
In most cases, a surveillance camera operator can pay attention to a maximum of 2 monitors at the same time and therefore it is possible that important events are not noticed at the appropriate time.
Due to the high cost to hire professionals in the area, companies end up ignoring this fact and require operators to view multiple monitors at the same time.
To have a cost-effective system, video analytics software has been developed to enable smarter event-based monitoring.
Event Analysis for CCTV
An event can be described as a specific action performed in front of a security camera and which can be recorded by the monitoring system.
A monitoring and recording system with embedded intelligence can be programmed to identify such events based on specific rules that must be adjusted according to the need of the application.
Motion detection, for instance, is nothing more than pixels that move on the screen when a person or object passes in front of the security camera.
Action and reaction events in intelligent video analytics systems
After identifying an event such as motion detection the monitoring system can react according to pre-established rules and perform specific tasks such as highlighting the camera image on the operator screen, triggering a siren, changing the positioning of a mobile camera in the local, send a message via email or SMS, etc.
In practice, an intelligent video analytic system is based on action and reaction and can be programmed in different ways to work with a lot of events.
Devices that work with intelligent video analytics
There are different types of devices that can work with intelligent video analytics, from analog professional cameras to standalone systems based on powerful servers, let's discuss each of these devices.
Professional analog camera
It is not common knowledge of the public that works with traditional analog systems that analog security cameras can do intelligent video analytics, although these analyzes are simpler when compared to the intelligence embedded in modern IP cameras.
Intelligent video analytics feature found in analog cameras are simple and limited due to the fact that there is no bidirectional and dynamic communication with the monitoring system, software or applications.
You can program a camera to identify motion in the image and display an alert message on the screen, show pixel tracking, or trigger an external device such as a siren or alarm system via I/O contacts (dry contacts) available on the own camera.
You can also find more advanced video analytics such as line crossing, entering and exiting areas, and a lot more that will be discussed later in this article.
The major limitation, however, lies in the fact that an analog security camera does not have the capability to send information from such events to the recording and monitoring system, nor can it send messages such as e-mails and alerts via SMS.
Intelligent video analytics embedded in professional IP cameras
IP cameras have long been used in modern monitoring recording systems since its invention by Axis Communications. IP cameras have caused a major revolution in the electronic security market due to the fact that it has embedded intelligence and the possibility of efficient communication with recorders, servers, software and applications over a traditional network of computers and the Internet.
The intelligent video analytics of an IP camera is more advanced compared to that of a traditional analog camera previously discussed and the possibility of communicating with external devices makes this function even more interesting.
A professional IP camera has an internal processor that is able to identify specific events in the image and trigger recording on a remote device such as an NVR (Network Video Recorder), and record to a memory card located in the camera itself, which enables the use of redundancy systems.
By performing intelligent video analytics and identifying specific events, an IP camera can send alerts via e-mail and SMS directly to specific individuals who may react according to some pre-established security protocol.
Based on events, an intelligent IP camera can also trigger presets and patrols on mobile cameras (PTZs) located at key points in an environment, as well as the possibility of triggering siren systems, lighting, alarms, access control or any other system that allows interaction via traditional universal TCP / IP network protocol.
Intelligent video analytics for Digital Recorders (DVRs)
A digital video recorder, popularly known as DVR converts analog signals to digital format to record on an internal hard drive and allows the images of the cameras to be shared via a traditional computer network.
In addition to recording, a DVR can also work with intelligent video analytics as it receives the video signals from the cameras in each of its channels.
Due to the fact that it is connected to a computer network, a DVR can send event information and intelligent video analysis to the software and applications, which allows the use of alerts via e-mails and SMS in the same way as an IP camera.
The intelligent video analytics available on a DVR depends on the make and model of each manufacturer, there are devices with more or fewer features and obviously, the more expensive professional models usually have more features available.
Intelligent video analytics for Network Video Recorders (NVRs)
Network video recorders, popularly known as NVRs, are used for recording management of IP cameras and thus have a natural way of communicating via a computer network, which allows them to send messages and trigger devices on the network.
An NVR can work with the intelligence embedded in its own processor and analyze the images it receives from each of the IP cameras or it can receive the result of the analyses carried out by the cameras themselves which relieve processing on the recorder.
It is usually more interesting to let the IP camera itself perform the intelligent video analytics and communicate with the NVR through pre-set events, but there are cases where the IP camera does not have the desired video analysis function or even has no type of intelligent analysis, and in these cases the recorder is fundamental to allow the use of intelligence in the recording and monitoring system.
Intelligent video analytics on servers
Software-specific servers are used for intelligent video analytics that is generally not available on IP cameras because of the device's processing limitation or simply because software companies need to provide more robust and scalable solutions.
Many of these companies provide solutions for large projects where it is critical to use systems that can make video analysis more complex and more accurate.
There are software developers who do not provide cameras or hardware for security projects because they are focused only on intelligent video analytics and integration with cameras and recording systems from different manufacturers.
Others, however, provide complete solutions with cameras, recorders and intelligent video analytics software that have powerful algorithms that can identify and classify events and interact with devices connected to the local network or the Internet.
The use of such intelligent software often requires the acquisition of high-throughput servers, although there is a trend in the market for distributed utilization of processing among IP cameras working in partnership with the server to achieve a final result with less centralized processing.
Types of Intelligent Video Analytics
There are different types of intelligent video analytics and here I will exemplify the most common and traditional found in the market of electronic surveillance.
This is one of the simplest analyzes found in cameras and monitoring recording systems, it is present in professional analog cameras, IP cameras, recorders, and professional software.
Note in the picture the menu of an IP camera that shows the option to enable Motion Detection, in which case the manufacturer separates this function from the general video analytics menu.
The principle is very simple when there is some kind of movement in front of the camera the pixels that form the image are changed and this is identified as motion detection by the device.
When a person passes in front of a surveillance camera it is possible to identify this pixel change in the image and generate an event for a device to start recording and record the "motion recording" information.
This recording can be performed on the IP camera itself when it has a memory card or on a recorder that is connected to the camera either through traditional coaxial cables in an analog system or through a network with switches.
A person who has a computer with monitoring software or a mobile phone with a specific application can be alerted about such a motion detection event.
Virtual line crossing
The names may vary depending on the brand of the equipment or software you are using, but in practice, the principle of operation is the same. This type of intelligent video analytics allows an event to be generated as soon as a person or object crosses a virtual line that must be pre-configured in the camera.
Virtual Line Crossing analytics allows you to generate events in different directions and in this way it is possible to know if a person or object is crossing a virtual line horizontally, vertically, from left to right or vice versa.
Note in the image below that a virtual line that has been configured to generate an event when a person or object crosses from left to right (see indicative arrow).
The virtual line also allows you to generate an event alert in both directions to simply identify when a person or object crosses a particular location.
See the following image with an example of a virtual line that identifies the vehicle passing through a given location, a green frame shows that there is movement, and triggers a line-crossing event. Note that there is a yellow frame around the image to indicate that there was an event.
There are many practical applications for this type of features such as the identification of vehicles traveling in forbidden directions or the movement of people entering or leaving a store.
Sensitivity, format, and object size adjustments allow for more accurate video analytics and reduce the number of possible errors that are common to this type of application.
See the following image for an example of specific area configuration for event detection
Imagine a person or animal crossing a virtual line in a region where the intention is to identify vehicles that go in the wrong direction, the fine-tuning of sensitivity should allow the reduction or if possible elimination of events that are an exception to the rule.
Such adjustments can be more or less efficient depending on the characteristics of the devices or software you are using, video analytics software is generally much more intelligent and accurate in event configuration and sensitivity adjustments.
This type of video analysis is also very common in IP cameras, DVRs and NVRs because it is very useful for situations where you want to know if a person or object has entered or left a specific area.
To activate such a function simply create a virtual frame that delimits an area where the object or person will enter or exit. Simply choose the desired option or both and adjust the detection sensitivity.
Note in the following image that an area on the left side has been selected and that people or objects entering the site are detected and an alert is generated automatically.
There are several types of applications for area entry and exit analytics such as identifying people and vehicles entering forbidden areas or removing objects on site.
Other types of video analysis
The intelligent video analytics presented in this article is basic but very efficient for the application in different situations where it is necessary to generate alert events.
There are other types of video analytics that are more advanced such as loitering detection, behavioral analysis, tracking, counting objects or people, character or face recognition, capturing and recognizing vehicle license plates and a lot more, however, they will be discussed in other more specific articles posted here on this blog.
There are several types of intelligent video analytics that can be used to assist the monitoring operator or even to alert people or activate external devices.
Each situation requires an analysis of what intelligence is to be implemented and there is usually a cost associated with the use of such technology, but the most basic features such as motion detection, virtual line crossing, and area input and output are very common in various IP camera and digital and network recorders.
Before acquiring a recording and monitoring system it is advisable to study what the design needs are and what makes and models of equipment can proportionally work with video analytics with the best cost-benefit.
In many cases, the basic technology that already comes embedded in the camera, recording devices or software is enough to meet the needs of the project.
Now that you already have the basic knowledge about intelligent video analytics on CCTV systems I suggest you put into practice what you have learned and continue to study other advanced features that are available in the market. Read the other articles here on the blog to keep yourself updated.
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 15+ experience in the surveillance industry. He has been traveling around 17 different countries to teach people how to design, and install CCTV systems.