Home » Physical Security » Video Surveillance

Industry View

Industry View: How to Avoid Five Common Misconceptions About Video Analytics

ObjectVideo's Alan Lipton offers suggestions for dispelling hype and finding value

By Alan Lipton, ObjectVideo

Page 2

The Truth: The performance of analytics systems is generally very acceptable in well-designed operating environments. If there are enough pixels and appropriate illumination to view an unobstructed area of interest, detection rates are usually very good. Most reputable manufacturers claim detection rates (in ideal conditions) of over 90% with very acceptable false alarm rates — as low as 1 per week per camera or even better.

Myth Three: Video analytics are "one size fits all". Customers often have trouble in the field when an analytics technology is taken outside of its comfort zone. There are many different video analytics products that are designed to do many different jobs in many different operational environments. Many vendors are even selling Video Motion Detection (VMD) as intelligent video analytics. In a largely unchanging, indoor environment, VMD might provide some value but in most realistic operational scenarios, there will be many false alarms caused by extraneous motion and the inability of VMD to distinguish between any motion and a real event of interest.

The Truth: Like any product, there are varying design specifications behind each video analytics application. Don't be fooled into thinking that a technology for counting people in a retail setting will work for perimeter protection or vice versa. Analytics products often make assumptions about camera orientation, distance to objects, numbers and types of objects, indoor vs. outdoor environments, illumination and many other factors that can affect performance. Make sure that whichever products and technologies you choose are appropriate for your operational environment and your particular mission.

Myth Four: Video analytics (or the necessary hardware) is not cost effective. Historically, video analytics solutions have required a serious capital outlay to get up and running. Per-camera license costs used to be very high; systems required large PC infrastructure footprints; and deployment, maintenance, and training costs were a constant financial question. In some cases this is still true! This is a very large expense for system that isn't going to replace a human guard-force any time soon.

The Truth: Video analytics technology, as it has matured in the marketplace, has become more of a value-added component of CCTV infrastructure rather than a stand-alone enterprise software solution. Today, it is possible to purchase true intelligent analytics capabilities as a feature of main-stream CCTV hardware components such as: cameras; IP video encoders; IP routers; and DVRs, NVRs, and video management platforms. Analytics features do add a premium to the prices of these devices but the price-tag is significantly lower than a stand-alone PC-based solution. Furthermore, more sophisticated customers of analytics have seen more value in analytics than simply replacing a guard-force. Customers are now reaping value from analytics by making their people, processes and infrastructure more cost-effective. This is being done by using analytics to help optimize bandwidth and storage of network video systems; or by using analytics to make forensic investigations faster and more personnel-efficient.