Video Analytics for Security Applications

From CCTV Information
Revision as of 23:29, 12 February 2011 by Gb003162 (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

What Is Video Analytics?

Video analytics is the use of sophisticated algorithms applied to a video stream to detect predefined situations and parameters. These can include but are not limited to the following.

  1. Motion Tracking - Following a moving object across the cameras view
  2. Object detection - Detect a moving object in the cameras view
  3. Object classification – Identify the type of object moving (Person,animal,Vehicle)
  4. Direction Flow – Identify the direction of a moving object
  5. Loitering – Alert if object stationary beyond a predefined time
  6. Left/Removed object – Alert if an item is left or removed from a predefined area
  7. Face detection – Detect and record faces
  8. People Counting – count people in and out of a defined area

There are many processes involved in developing a video analytics platform but the main areas are, obtain the video data this can be from either an analogue camera or an IP camera it could also be with some systems from a video file (avi, wmv). The video images are then passed through image extraction, feature and edge detection algorithms. These will check the image for recognisable details, compares the foreground to the background and differentiates between them by only detecting foreground objects.

At this point the systems have enough information to make intelligent decisions based on the pre-set rules and parameters set in the software. It can then either ignore or raise alarm depending on these rules.

Video Analytics in the real world

This is where VA differs significantly from other methods of detection i.e Motion detection, Infra red and seismic detectors. VA uses recognition and classification to decide whether the object is a threat or not. For example, normal detectors would pick up the presence of wildlife and trigger an alert as all it knows is that there is something moving whereas video analytics could detect the object and classify it as not a person and ignore it. The same applies for other everyday situations such as heat extraction units and even swaying bushes. These would most likely activate an infra-red sensor or standard motion detection systems but would be ignored by VA system. This level of detection massively reduces false alarms. The other huge advantage is in control room environments where typically staff monitor tens if not hundreds of cameras, the VA system can be used to analyze all the video feeds and alert the control room operator only when a specified condition has been met. This therefore improves the efficiency of the control room operatives in handling genuine alarms and dismissing false alarms.

Are Video Analytics difficult to Install?

In general they are no more difficult to install than a conventional CCTV system and the normal rules apply to lighting, camera view, camera location considerations. Be aware that the analytics system can only detect inside the cameras field of view and there has to be enough light for the image to be clear enough for the system to work reliably the required image size and light level will vary depending on manufacturer.

There are generally two types of Video analytics systems

  1. Edge based units. These use analytics either embedded into the camera or on a standalone decoder sited near the camera.
  2. Mass centralised servers. These are units which can connect to existing or new analogue or IP cameras and do all the analytics in the server.

Which type you use for the installation will depend on your personal site requirements.

What else can Video analytics do?

Along with all the obvious security uses video analytics is being developed constantly and new features create new opportunities in different market places. Areas such as marketing and retail can also benefit from a VA systems as it can track how many people enter a building or area, how long people stay in certain areas. It can also monitor queue lengths, queue times, provide heat maps and show customer flow. All this information allows the end user to have a better understanding of how their business operates and provides measurable performance.

This article was supplied by Business Insight 3.

If you wish to discuss how to make use of video analytics, Contact us to speak to a consultant.