Analyzing Video Analytics

Recently, I had the opportunity to sit down and listen to some of the top video content analysis (VCA) experts in the country. Many of the discussions centered on the deployment of video analytics (VA), some related to security and others not.

If you haven’t noticed, this technology is now getting the attention of many communities and businesses all over the country as it is being deployed in urban areas, transportation hubs and by major corporations.

If you are a security value-added reseller (VAR), integrator or end user who has been sitting on the sidelines with reference to VA, now is the time to get into the game as the tipping point has begun. The technology in many instances is moving past the early adopter stages and is proving itself a valuable and cost-effective tool in the fight against crime.

Managing Customer Expectations

As with many new and glamorous technologies there is a lot of hype, glitz and glamour surrounding VA. I appreciated the conservative approach these experts used to explain some of their case studies.

It is important, as with any project, to identity your customer’s or prospect’s expectations. My choice has always been to get defined expectations in writing. Sounds simple, but again, missed by many.

One of the experts cited an example of a large municipality that implemented an urban IP video surveillance system with the reasonable expectations of viewing five frames per second (fps). However, due to bandwidth problems, the customer ended up with one frame every 10 seconds.

As it was emphasized by Joe Heinzen, president of Convergence Solutions, an experienced distributor of VA solutions, “While many VA solutions work today, it is important to develop a good, solid relationship with your vendors.”

Analytics Vs. Motion Detection

According to the experts, another key area is confusion between VA, sometimes referred to as VCA, with video motion detection (VMD). If there is any correlation at all it would be that VMD can be considered a type of VA. Along this line of reasoning, VA also could be facial recognition, automated number (license) plate recognition (ANPR) and vision-based inspection.

The big concern here is that VA often includes higher levels of analysis and algorithms than VMD, thereby making it considerably more reliable and less prone to false alarms. There are basically three tests to help decide if a video system is intelligent: It must have some form of advanced artificial intelligence; it must be able to provide some form of metadata; and it must be able to differentiate between foreground and background objects.

Going to the ‘Edge’

Some of you old-timers may remember the early evolution of access control. In many ways VA’s progress is similar. During the early years of access control everyone planned to have multiple card readers with one central controlling point. However, early on it was found that if personnel entered many doors at the same time such as coming in to work, delays in communications caused major problems. It was quickly decided that taking reader controllers to the reader locations was a necessary solution. This was the first march to the “edge.”

If you enjoyed this article and want to receive more valuable industry content like this, click here to sign up for our FREE digital newsletters!

Security Is Our Business, Too

For professionals who recommend, buy and install all types of electronic security equipment, a free subscription to Commercial Integrator + Security Sales & Integration is like having a consultant on call. You’ll find an ideal balance of technology and business coverage, with installation tips and techniques for products and updates on how to add to your bottom line.

A FREE subscription to the top resource for security and integration industry will prove to be invaluable.

Subscribe Today!

Get Our Newsletters