So each time a unique individual comes in my location, I can link them up with their name, order, etc, and track who's who.
Dahua Smart features claim facial recognition, which I know isn't what it sounds like... but is a scaled back expectation like the above within the realm of possible?
That sounds reasonable. It would mainly require an integration between the facial recognition / surveillance system with whatever CRM or customer management system you use.
What application would you want to display the customer details in? In the surveillance system or show the face in the CRM system?
I would do it any way it would cooperate. Either in Filemaker, Milestone or Dahua NVR (whichever I decide on - as someone has compelled me sufficiently to experiment with it) :) ...
Do you think there's an API?
The goal (and distinction of what makes it worth doing or not) ... is that it would notify me wit conditional reactions to categories of people who've been identified, that return.
That would make it good for businesses that have had shoplifters, difficult customers (I feel like making a counter program to yelp called DCDB.com - (Difficult Customer DataBase). Bars that have 86'd people, casinos who with card counters or high rollers... and that last group is the group that makes me feel like this sort of thing exists, and has, for years and years.
Thanks
Do not put a lot of weight into Dahua Analytics. We have been using their products for years, and recently have been playing with certain features like line crossing, abandoned item etc.
Some of the core stuff works, but its not polished and completely finished.
I figured, because they sound like pretty hard things to do, especially as the ratio of pixels per foot decrease, and the depth and area square when you double the distance. I would however, imagine that the shorter the distance to the camera (higher PPF) will make it easier.
That said, I'd bet there are servers that are far more capable of analytics than a device that can't even use a fan to cool itself for the obvious reason that cameras have to be designed with the concern of fan noise attracting attention to it.
Kaby Lake Future
That said, I REALLY CANNOT wait for the Kaby lake (and on) Intel series... which I have to think the features will permeate all devices as they get down to the $10 range in a year or two for the bottom end... as their $39 CPU:
G3930 has built in decoding of 4096x2304@60Hz (FPS I assume) ...
which contrary to Axis original reason for apprehension of adoption, is going to be revolutionary for computers.
Off to Fry's to buy the cheapest smallest Kaby Lake workstation to use as a decoding station and see how it works. :)
...especially as the ratio of pixels per foot decrease, and the depth and area square when you double the distance.
Just to note that PPF is a linear measurement, therefore if the distance doubles, the PPF only doubly decreases.
I'm inclined to disagree; this is a physics principle with everything that "radiates" - i.e., since the only useful data from a camera is the azimuth that hits a useful subject, it's "radiation" - and is subject to the inverse square law. Even without doing the math, I'm tempted to assert that. But, that is provisional, and I'm happy to be wrong as well as research it later. Short on time ATM.
Thanks. :)
I'm inclined to disagree... since the only useful data from a camera is the azimuth that hits a useful subject, it's "radiation" - and is subject to the inverse square law.
Does the inverse square law apply to cameras? Yes. However, PPF is just a simple calculation of how many pixels are representing how many linear horizontal feet.
So if the scene width doubles, then the PPF gets halved.
On the other hand, Pixels Per Square Feet, is an interesting concept...
actually - yeah... now that I think of it, it is just the function of a triangle, not the area. Good point.
But yeah, the area would be the inverse square law. I lay corrected. :)
http://www.hertasecurity.com/en
Ask questions and get answers to your physical security questions from IPVM team members and fellow subscribers.
