Specialized Video Analytics

The core capabilities of a Video Analytic system have been described in the previous chapter along with their iQ levels. In this chapter, you will learn how some of these products have been extended for use in Smart Cities. Using its software building blocks, iOmniscient has created new applications that are particularly relevant to the needs of various stakeholders.

As an example, one often has to manage queues in cities. iOmniscient offers a queue management product that uses a combination of the capabilities of the counting products, the crowd management product and the Face Recognition product to provide the user with information, such as the average waiting time in the queue, the proportion of people who leave the queue before they get to the end of it and so on.

The main subjects for surveillance in a Smart City are people, vehicles and events. Hence, the challenge in a Smart City system is to understand how people and vehicles behave and the types of other events that may occur. In this chapter we provide a few examples to demonstrate the possibilities.

Back to top


Applications around People

Specialized Analytics - People * IQ-Smart City

While occasionally there may be a requirement to protect perimeters and prevent intruders, most Smart City applications approach more complex human behavior. Some interesting applications are described below.

Control of Access to Restricted Areas

Accesses to important public buildings (and car parks) may need to be controlled. Standard access control systems can be easily defeated. Access cards can be used by unauthorized personnel. Or an unauthorized person may tailgate behind an authorized person.

In such situations, iOmniscient's Face Recognition in a crowd system can be used to authenticate that the holder of the access card is indeed its owner. Tailgating can be detected and the faces of the culprits can be recorded using iOmniscient's Face Detection system. Similar systems exist for controlling the access of vehicles.

Vandalism and Graffiti Detection

The system is able to detect vandalism and graffiti in a crowded scene. Often when graffiti is being sprayed, the culprits are performing their art in open view. There may be passersby on the street and the intent would be to ignore these people and to only raise an alarm when new graffiti is seen. This can be achieved by using the iQ-140 system. The same is true for vandalism, such as the breaking of glass panes or the damage to street furniture.

Detection of ATM Skimming Devices

Identity thieves have developed various devices to skim information off credit cards that people insert into ATMs. These can be very well camouflaged and very difficult to see. iOmniscient's anti-skimming device system, based on iQ-180, can be used to detect such devices. While fraud prevention from ATMs is of particular interest to banks, the catching of identity thieves is usually important to law enforcement authorities and therefore, both tend to have an interest in it.

Loitering at Night/Soliciting

In some jurisdictions, law enforcement agencies have an interest in detecting people who may be loitering in an area. They may be car thieves who are checking out vehicles to break into or others who may be soliciting for business. Such behavior can be detected using an iQ-115 system.

Gathering of Groups

At times, it may be necessary to know if crowds of a certain size are gathering. This is often the case where a meeting area is only designed to hold a certain number of people.

The iQ-120 system can provide information about the number of people in a given area and it can raise alarms if the number exceeds a given threshold.

Fighting

Fighting is difficult to detect reliably using Video Analytics because the body language of people coming together from different communities can be very different. In some communities people may get very close and hug each other. Therefore, attempting to detect fighting can result in a system prone to false alarms based on different social norms. In the next chapter, we will discuss Audio Analytics which can complement Video Analytics. By determining that people in the vicinity have noticeably raised their voices the system can be used to detect potential fighting.

Management of Public Spaces

In addition to detecting slips and falls, a key application for the management of public spaces is counting people as they move from one area to another, for example, in a shopping mall. Understanding traffic flows of people is critical for the running of the shopping precinct. It allows authorities to determine how many people travel along a particular route on which retail rentals can be based on. The retail shops themselves can use traffic counts to determine where to put the chocolates and where to place the milk.

The most important aspect that differentiates “counting” from other iOmniscient's Security Specialized Video Analytics Capabilities for a Smart City type applications is that counting can help the user to make money, whereas all forms of security only result in the spending of money. Marketing departments can find it very easy to justify spending money on counting while security departments may struggle to get the money they need. However, after a crisis, management will often spend on security AFTER they have had a major incident. In a city, the local council may wish to count how many people walk through (or drive through) a particular direction so that they can sell advertising space. Alternatively, they may use such information to provide improved services. For instance, if a large number of people attempts to cross a road at a certain point, it may suggest that they need a foot bridge or a traffic light there.

Counting is usually a statistical activity. Whereas it is valuable to have a count as accurate as possible, a small variation is usually not relevant in terms of the information that can be derived from the data.

Counting systems can yield some very interesting information. They can provide information on the routes that people take as they meander through a public space. It is possible to compare traffic information with comparable, previous periods. Using iOmniscient's Face Recognition system, it is also possible to understand the demographics of a group of people. The system can provide information on their gender, age and ethnic background which may be used for marketing but also for security purposes. If, for instance, a young lady or a very old person is walking down a very lonely alley, the system can raise an alarm and alert the police to be keep an eye out for them even though their identity is not known.

Crowd Management

In busy cities, groups of people will inevitably gather and the authorities will want to understand the dynamics of these crowds. They may wish to know when small groups (say of 4 to 5 people) get together. Or they may wish to understand the total number of people in a particular area They may also wish to know if known trouble makers or other persons of interest have intermingled with the crowd. Having accurate information on how a crowd has come together and how it moves can provide the authorities with the means of maintaining orderliness and peace so that citizens can partake of major events without danger.

Tracking People

When large numbers of people have to be processed, many queues may need to be formed. iOmniscient's Queue Management System calculates average waiting times and advises officers if more support is required to process the queue.

There are several methods of determining the length of a queue. One technique involves the use of the counting system as a base. In this case, one counts how many people enter and leave the queue and, based on this, their average waiting time can be established. An alternative method is to achieve the same result using Face Recognition to determine how long it takes a particular individual to go from Point A to Point B.

Queue Management

In addition to detecting slips and falls, a key application for the management of public spaces is counting people as they move from one area to another, for example, in a shopping mall. Understanding people traffic flows is critical for the running of the shopping precinct. It allows authorities to determine how many people travel along a particular route on which retail rentals can be based. The retail shops themselves can use traffic counts to determine where to put the chocolates and where to place the milk.

The most important aspect that differentiates “counting” from other security Specialized iOmniscient Video Analytics Capabilities for a Smart City type applications is that counting can help the user make money whereas all forms of security only result in the spending of money. Marketing departments can find it very easy to justify spending money on counting while security departments may struggle to get the money they need (except after a crisis - management will often spend on security AFTER they have had a major incident). In a city, the local council may wish to count how many people walk through (or drive through) in a particular direction so that they can sell advertising space. Alternatively they may use such information to provide improved services. For instance if a large number of people attempt to cross a road at a certain point it may suggest that they need a foot bridge or a traffic light there.

Counting is usually a statistical activity. Whereas it is valuable to have as accurate a count as possible a small variation is usually not relevant in terms of the information that can be derived from the data.

Counting systems can yield some very interesting information. They can provide information on the routes that people take as they meander through a public space. It is possible to compare traffic information with comparable, previous periods. Using iOmniscient's Face Recognition system it is also possible to understand the demographics of a group of people. The system can provide information on their gender, age and ethnic background which may be used for marketing but also for security purposes. If for instance a young lady or a very old person is walking down a very lonely alley, the system can raise an alarm and alert the police to be keep an eye out for them even though their identity is not known.

Fever Check

When there is a health crisis, e.g., the SARS crisis, it is important to know if sick and potentially contagious individuals are on the street. iOmniscient's automated Fever Check system can be used for automatically detecting if a person has fever.


Applications around vehicles

Specialized Analytics - Vehicles * IQ-Smart City

Speed Detection

Speeding vehicles are always a hazard and occasionally they may be the precursor of a dangerous event. iOmniscient can detect the speed of vehicles and raise an alarm if a vehicle is detected exceeding certain authorized speed limits. Traffic speed can be calculated using several different methods. The speed of every vehicle can be measured individually. The average speed of the traffic can be calculated. The time that it takes a vehicle to go from Point A to Point B can be calculated to provide its average speed which can be important to avoid a situation where vehicles only slow down at “Speed Cameras”.

Automated infringement notice generated by using License Plate Recognition due to speeding violation - administered without human intervention Tracking vehicles

Just as humans can be tracked using Face Recognition, vehicles can be tracked using their License Plate. If vehicle needs to be tracked, the system can locate all the cameras on which that vehicle has been spotted. If there is good coverage of the roads, it should be possible to know the exact route that a vehicle has taken.

Back to top


Applications around events

Specialized Analytics - Vehicles * IQ-Smart City

An example of an event without human or vehicle involvement is a fire.

Smoke and Fire Detection

Standard smoke and fire detection systems cannot always be used in busy city environments, especially if they involve open air situations. iOmniscient's video based smoke and fire detection system can be used for very early detection of such events.

There are many inexpensive devices available for detecting smoke in small confined spaces. However, these do not work well in large cavernous places or in the open air. For such situations, the use of video to analyze the situation provides a faster and more accurate alternative.