Video Analytics Market Guide and Favorities 2014

Analytics, like good Licence Plate reco can't be done with good results " sometimes" if your are a SI. This should be practiced regularly and become also a strongpoint for sales to enter new markets and create new needs. (and not only install cameras and record) Some installers seling previously low end systems are moving to analytic and don't fight anymore on price per camera, but on "effective solution" coupled to alarm and Access control.

All analytic / LPR installations done only to answer an opportunity but without real knowledge of False alarms in video are often very disappointing.

As a lot of Indoor Motion detection are often not very well configured and generates false positives or false negative (worst) especialy at night, you can imagine that calibrating a deep FOV with shadows, Light reflects, Wind, Snow/rain, Bugs and vibrations can't be improvised and need a good testing method to reach good performances. (anyhow powerful the camera are, or the analytic server...)

Great issue, which brings up a couple of thoughts.

First, does IPVM have a simple catalog of video analytics? I mean, a simple list of "stuff" video analytics is supposed to do. For example:

Breaking a line

Crossign a line in a particular direction

Entering a box or geometric shape

Leaving a box or geometric shape

... all the way to ...

Recognizing celebrities

Recognizing that my very demaning Aunt Telula is visiting and notifying all of the family sufficiently early that we can stow all signs of activity and convincingly pretend we're not home by the time she knocks.

That might be an interesting list.

----

Second, I wonder if you have any sense of what people wish video analytics could do for them. This might be an interesting open survey, not just to industry professionals.

For example, I imagine that those parent whose kids were playing in the nearby field would like to have known that a police cruiser had stopped nearby and that a man was approaching their children. I think the parents would have liked to have got to their children first to have headed off potential problems.

If the costs were marginal, most parents wouldn't mind having a catalog of comings and goings at their home.

If line crossing is the best they've got, that may be swell for a few specialized applications, but we've had pretty reliable and inexpensive optical trip wires for a while now.

I suppose package leave-behind helps us when the UPS man doesn't ring the bell, but that doesn't ring my bell. Much.

Unbounded by today's capabilties, would people speculate on what they wish a video analytics "killer app" might do for them?

Let me offer an example. Currently lawn sprinkers typically use timed zones to meter the amount of watering, sometimes tempered by whether it has rained recently. Lawn sprinklers are somewhat expensive to install owing to the excavation requirement. For less money, one could easily imagine a ring of 1/2" PVC around the house nestled just below the gutters. Trainable nozzles could sweep watering areas evenly. However, when integrated with video surveillance, the system could provide feedback to track where the water really landed in order to ensure more even coverage. In directable spray systems, or even by duration in zoned timed systems, if there is any optically detectable under- or over-watering indications, the system should easily compensate. For example, in our real-world buried sprinkler system we know from experience that over time some marginal areas start to show yellowing or browning effects. Some sprinkler systems allow selective application of fertilizers within the water, which could also be selectable to match areas that are deficient in some way if only that were also reliably detectable optically. Beyond that, I've noticed the neighbor carries a pitcher of water whenever they walk their dog, diluting the dog's natural fertilizer to prevent burning. Our lawn shows brown burns where our dog did his business, but theirs is a lush green. Couldn't the video tracker note and remember doggie pause locations for dilution treatement when everyone is clear? These sorts of integrated tools might take some sophistication to implement fully, but the sky is the limit. PS from Hollywood of course the option would then exist to use a directed spray to chase off deer eating your lillies, or unauthorized dogs using your lawn for a potty, or unauthorized intruders based upon some programmable or detectable characteristic or behavoir. Costs overall should be comparable (worst case) to buried sprinker systems and "normal" (whatever that means) video perimeter monitoring, but the synergy could provide a greater capability when integrated. While all this may not be a killer app, it may point to ways in which greater integration of existing systems could add meaningful value. PPS think of the kind of dough some people are willing to pay for well kept lawns.

This example might have been more "nice to have" than "killer app," but it might get the mind thinking about what sorts of outcomes might be possible if we move beyond trip wires and abandoned packages. Within a hodgepodge of brainstorming from every quarter, who knows? A few might stand out as really deserving of the challenges to develop and market such a really valuable capability.

Excellent !

John ? Suggestion for a new department ?

We use analytics extensively to protect critical infrastructure. I have found the success rate for the use of analytics depend on the following:

• Lighting (LED or Metal halide)
• Mounting height (manufactures specification)
• Camera / lenses selection (never exceed manufacturers minimum pixels on target specifications)
• Proper installation (manufactures have trained the technicians)
• Selection of the right product for the job (sometimes we use a Videofied system Vs. a high end manufactures product)
• Analytics on the edge over VMS based systems
• Proper system design and installation (you get what you specify and inspect)

While we do have some dual technology systems I can say with certainty that we could not protect our assets as effectively without the use of analytics. We also have internal system where we review alarm counts weekly and adjust our systems accordingly. In addition to the alarm count review we also partner with our Security Managers and Business unit leads to effect positive change on employee behavior patterns.

I suppose John, that the camera equiped with analytic require a much more regular preventive maintenance than ordinary camera (FOV, bugs, dust ..etc) and that also some regular performance testing and stress testing should be regularly performed to be sure that the security level is kept. Probably far from what we see on ordinary installations where maintenance isn't even mentionned. True ?

John raised some important points.

1. Following the manufacturer guidelines is a crucial element in a successful deployment of Video Analytics.
2. Having a factory trained tech. for some odd reason RTFM is an acronym that the LOL generations don't seem to be familiar with. Nor with the actual practice of reading the manual.

Video Analytics has suffered immensely from industry players who have promised the world and failed to deliver. those who get burnt in that manner will never trust the solution again.

Before approaching any new project I ask the following questions:

• People are not allowed 3 meters from the fence line
• Provide an alert when people cross the grassy area

• No one should walk within the fence after midnight
• This area should be sterile of human. Access control event will shunt the detection zone when maintenance team arrive

FALSE ALARMS are a part of any detection system, PIR's, laser beams, Microwaves, fence shakers, etc..

A good outdoor analytics has one main advantage over those: It will always provide a graphical OSD of the alarm trigger. So a false alarm is typically easier to ignore or escalate.

THERMAL CAMERAS are a great companion for Video Analytics as they eliminate the major cause of false alarms: Shadows, head lights, etc. Thermal camera also provide a longer range of detection, typically 3 times or more of similar optics on a day/night camera.

In a recent comparison of cost, we have concluded that for protecting an outdoor perimeter Thermal Cameras end up being cheaper simply for the added detection distance and lower count of cameras required. Once you take into a account the cost savings on infrastructure: Poles, wiring, lighting. Thermal cameras end up being much cheaper! In the long run, when you add savings on electric bill for lighting, and dealing with the environmental groups over the light pollution...Thermal is surely the way to go.

In summary, Video Analytics is a useful solution for many applications, but it has been traditionally specified for the wrong applications and for the wrong reasons.

Managing expectations is the one most important rules of thumb I would advise for anyone selling/installing Video Analytics.

Have done some tests with IPsotek, especially in an outdoor environment using both Day/Night and Thermal cameras. The results were satisfactory.