Analytics under surveillance...

Video analytics will detect ‘suspicious movements’ made by people walking along the street... Analytics will pick out an offender from a sea of faces... Analytics will detect terrorists... Just a few of the misconceptions about video analytics ‘doing the rounds’ at present. Was there ever a security technology that was so “over-promised and under-delivered”?

The reality is that video analytics are still very much in their infancy. Expectation management is at the heart of the issue here. In other words, being realistic with end users about what can and cannot be achieved.

The fundamental problem is that, as human beings, we carry out tasks without even thinking about them. We read licence plates and recognise faces in a totally subconscious way. It may have taken us years of learning during childhood to acquire these skills, but the truth is that we take them for granted.

Computers, on the other hand, lack even the basics of visual intelligence. They can perform some video analytics functions pretty reliably, but all-too-often as a result of severely constraining the application. Qualification is everything. Setting the end user’s expectations at a realistic level is absolutely vital.

Analytics in today’s world

Licence plate recognition has been around for a long time and is well proven. However, it’s still not 100% accurate. Facial recognition is notoriously difficult to perform on a reliable basis, such systems being extremely easy to fool by the use of disguise. For them to work with any degree of accuracy, an excellent headshot of the subject is a base requirement.

However, there are some bread-and-butter analytics functions that can be performed reasonably well today. Motion detection is the simplest, most basic form of analytics. Many manufacturers support it, but very few systems achieve sufficiently low ‘false alarm’ rates to be genuinely useable. A system generating anything in excess of 20 false alarms on any one night is rendered ineffective. The motion detection function may even be switched off just to prevent the operators from being ‘drowned’ by alarms!

Congestion detection has evolved from basic motion detection. When the density of humans or cars reaches a certain level an alarm is triggered. Counter-flow, meanwhile, looks for objects moving ‘against the flow’ and, it must be said, is highly valued by security managers in airports, for example.

Virtual tripwire is also a refinement of motion detection in that it triggers an alarm when someone (or something) ‘breaks’ a line drawn in the video image. This is useful in applications such as those large areas playing host to ‘No Go Zones’ (factories and industrial plants being good examples). Individuals are allowed to move around in ‘free areas’, but the system will alarm as soon as any one person moves outside of the jurisdictional barriers (see panel ‘Video analytics: the A to Z for security managers...’).

Taking all of the analytics applications mentioned thus far together, camera positioning, lens selection and lighting for the surveillance systems are going to be critical. Just changing the camera position can easily improve analytic performance by an order of magnitude. For example, in counter-flow detection the algorithm has a far easier job if the camera is positioned pointing downwards such that it sees the area in plan view.

It’s a competitive marketplace

There are literally hundreds of companies touting the fact that they can offer analytics software. As a manufacturer of complete IP-based video solutions, we are approached by (on average) one analytics provider per week, all of them asking us to integrate their product with our own IP video management platform.

Why so many? Simple. Any small software company can develop a suite of video analytics by buying a frame grabber and a powerful PC, and then writing some software to go with it. These are then sold as separate, stand-alone systems that sit ‘next to’ the main CCTV system. Video is split from the matrix and fed into the analytics system. This has the limitation that it’s not truly integrated with the operation of the main surveillance system and, therefore, will deliver only limited benefits.

IP-based video management systems provide the ideal platform for powerful analytics to be completely integrated into the system, making them a core and integral part of its operation. Leading IP video solutions support analytics that can be performed in two fundamental operational modes: live (to detect events when they occur) and post-processing (to test scenarios on recorded footage).

The optimum place to locate live analytics is obviously at the camera, as it’s the only truly scaleable solution and, in addition, doesn’t use up valuable network bandwidth. A camera with built-in analytics is able to monitor scene activity and transmit only on specified events (for example, an individual moving the wrong way through airport security).

The optimum place for locating post-processing analytics is obviously on a central server so that recorded video can be searched many times with different parameters.

Computers do best what they are good at locating possible security events while humans should stick to their best task. In other words, verifying them.

Video analytics: the A to Z for practising security managers, facilities professionals, specifiers and consultants

Abandoned object detection

Live: Used for alarm generation when an object has been left in a busy scene (such as a suitcase in an airport or railway station). This feature is a key component in the timely management of dangerous situations. The functionality here can also be used to detect illegal parking, or vehicles remaining for too long in certain zones.

Recorded: Abandoned object detection can also be used to search recordings for events like parking violations/blocked roadways.

Counter-flow

Live: Counter-flow is available to alert the user to a person or vehicle moving in an unauthorised direction (for example, an individual who is moving against the permissible flow in an airport’s immigration or customs area), or a vehicle travelling in the wrong direction on a dual carriageway.

Recorded: Counter-flow analyses can help optimise crowd control in public areas (such as the London Underground and mainline overground stations).

Motion detection

Live: Motion detection can be deployed to alert security personnel of unauthorised entry, and of potentially dangerous situations (for instance, if a member of staff is entering a hazardous area without protective clothing).

Recorded: Users can define specific areas of interest in a scene and search automatically through a recording to identify and view any significant motion that occurred during the recording period. This is extremely useful when searching for security events in corridors, on staircases and along walkways, etc during quieter periods. Can be fine-tuned using parameters such as object size and sensitivity.

Shape-based detection/object tracking

Live: Shape-based detection/object tracking may be used in a variety of applications. It can alert CCTV operators when a high-sided vehicle approaches a low bridge, for example. Alternatively, it may be used to distinguish between an animal approaching a boundary fence and an intruder.

Recorded: In recorded video footage, this feature can be used to analyse the types of vehicle travelling on certain types of road and at what time of the day (or night).

Theft detection

Live: Museum mode can be used to detect theft, such as the removal of a painting from the wall of an art gallery. In this mode, sensitivity is configurable while moving objects in the foreground will be ignored.

Recorded: Theft detection can also be used when reviewing recorded footage (for instance in a warehouse or stockroom). Able to rapidly identify when a particular item was either moved or removed from the scene altogether.

Virtual tripwire

Live: With a virtual tripwire set alongside a railway track, the hard shoulder of a motorway, a building’s perimeter or around a temporarily stationary asset, for example, the operator will be informed when that tripwire is breached. Since the system ‘understands’ direction, alarm discrimination based on the direction of approach is rendered a possibility.

Recorded: A virtual tripwire can also be placed on the entrance to a building or designated car parking area to review how many people or vehicles enter.

Congestion detection

Live: Congestion detection is deployed to alert a user in the event of a build-up of congestion in a given area of interest (for example railway station platforms, motorway exit/entry slip roads, EPOS queues, etc). This helps to initiate timely intervention and prevents an undesirable situation from worsening.

Recorded: Congestion detection can also be used to provide statistics for staff planning and marketing functions. For example, it can detect when a shopping centre is at its busiest, or when supermarket queues begin to build up. This aids staff deployment.