Can Search Improve Transportation Video Surveillance?

Published Oct 07, 2009 00:00 AM
PUBLIC - This article does not require an IPVM subscription. Feel free to share.

In an October 2009 article, 3VR contends that improving the ability to search video would significantly enhance transportation security.

Accomplishing this faces two main barriers:

  1. Most advanced video surveillance search functionalities are integrated into an overall video management system, meaning that when you buy their search, you also get their video recording. Of course, most transportation systems already use a variety of other video management systems, requiring the user either to replace or overlay this searchable system (disruptive and expensive). This is a fact for life for most of the systems including 3VR themselves and other advanced search providers such as VideoIQ and Bosch. The only company building their search as a plug-in to existing VMS systems is AgentVi though this was only announced last month.
  2. Searching over large cameras and time frames can return a nearly unusable result set. For instance, in our July 2009 test of VideoIQ's color/object search, queries for 'red cars' missed everything during low light or night times and returned hundreds of objects per day. Though it worked basically as 'advertised', the sheer volume of objects and the natural light changes clearly impacted the effectiveness at large scale. While it almost certainly will be better than nothing, it's likely to find thousands of needles in a haystack.
The other broader challenge is that most security organizations prioritize real-time alerts over searching - simply because stopping a crime is worth dramatically more than solving one. With that type of focus and practical challenges in implementing search, the days of widespread 'googling for video surveillance' could be far in the future.