Home > Intelligent Video > This Video Analytics Podcast is a good primer on the technology and the challenges it addresses

This Video Analytics Podcast is a good primer on the technology and the challenges it addresses

For this podcast, I interviewed one manufacturer, several integrators and some end users.  We discuss what video analytics are and what they aren’t, how to extract value from analytics, and how to screw it up.  In the end, we get a clear view of the current state and usability of video analytics with insights to the future.

Click Here  Download securitydreamer_on_video_analytics.mp3

Categories: Intelligent Video
  1. September 25, 2008 at 9:02 am

    Always good stuff, love the tunes as well. I have the following comments:
    So video analytics is NOT software, at the end of the day its a system, camera, lenses, lighting, hardware, software, databases, operators, etc.
    Its NOT friggin new, one of my previous companies put a license plate reader that worked in the Dartford Tunnel in London in 1979. Arggh, can people please stop thinking this is something new. What is new is Moore’s law allowing this to be done in a camera with a DSP or in software as opposed to the PDP 8 with 22 special purpose hardware modules we had to use. And this directly relates to the solution cost.
    It’s not changes in pixels it’s their relationships to one another or of groups of pixels to other groups of pixels typically in the same location in the image, and their comparison to a baseline.
    You don’t achieve the same effect with a virtual fence as you do with a real one, the deterrence is very different.
    It doesn’t matter whether someone is looking at a monitor or they are simply manning a guard post, people eventually get tired and lose concentration. So machine vision (video analytics) can provide improvements pretty clearly when compared to human vision. Its not about acuity its about concentration and consistency.
    CCTV can be event driven by combining other sensors. For the example given about cars in a tunnel if you detect a car you can switch to the camera, you don’t even need to be capturing video unless you get a detection. Video analytics must be able to provide benefits when compared to this. As an example combine video with radar. This allows you to have a single camera with a strong and fast lens combined with PTZ that can cover a wide area.
    Analytics’ effectiveness is driven by the resolution of the camera and the size, type and “contrast” of the object that you are trying to detect (independent of the complexity of the problem).
    There was a very cool concept presented. The idea of having 24 channels of analytics and then apply where needed presents a very powerful ability to adapt intelligent surveillance to the changing threat profile. Those that can do this will find their sales and value to their customers grow, notwithstanding the required expertise.
    Another good comment was pointing out the ability to use SDKs. The thing that drove sales in the machine vision market was the ability of vendors to find people who wanted to use the solution on an OEM basis. So to your point, 2.0 comes when its moves forward as a OEM product incorporated into a security suite.
    And YES, its about managing customer expectations!!!

  2. September 25, 2008 at 12:23 pm

    Awesome comment. Thanks, Sal!

  3. September 25, 2008 at 11:36 pm

    Very useful podcast Steve, thank you very much!
    In my opinion, video analytics today is still 90% hype and useful in only a relatively small number of real world applications. No doubt that intelligent video motion detection systems have their uses and the systems today are far better than they were 20 years ago. Beyond that, the number of other useful video analytics applications that are cost effective in a commercial setting can be counted on one hand.
    The real issue here is that too many clients see video surveillance as a “magic bullet” that will solve all of their security problems.
    Instead of developing a comprehensive and balanced security program based upon a formal risk assessment, they try to buy their way out of their security problems by installing cameras: in some cases, hundreds or even thousands of cameras.
    They then find they can’t watch all the cameras or even begin to manage the recorded images which now require many terabytes of storage. To solve this problem, they again turn to technology such as video analytics to make their camera system “more effective”.
    The truth is, cameras weren’t the solution to begin with, and providing better analysis and processing of the information isn’t going to make things any better, either.
    Don’t get me wrong; video surveillance is a useful tool, and when properly applied, can be of great benefit. However it is my opinion that video surveillance is probably one of the most overprescribed “solutions” in security today.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: