WestPoint · Cameras · Avigilon
Avigilon does not sell cameras. It sells a system, and the cameras go inside.
North American, today part of Motorola Solutions. It came from video software and from search, not from optics, and that explains where it really performs and what commitment you sign when you choose it.
What kind of manufacturer it is.
The other way round from usual: it is not a camera maker that added a program, it is a system that makes its own eyes.
It started making video software and analytics, and the cameras came afterwards to feed it properly. That is why the equipment is thought out so that the system gets something out of what it records, and not just so that the image looks nice on a screen.
It puts two very specific things first: that the drive lasts and that searching is fast. Its way of compressing cuts a lot of storage without taking away the detail of what moves, and its search lets you point at a person or a car and ask for other similar appearances. With a hundred cameras, the first is money and the second is hours for whoever reviews.
What it sacrifices is neutrality. Outside its own system the camera works with somebody else's software, but it stays a good, expensive camera: what makes it different lives in the other half. And part of the range goes towards the cloud with a subscription, which is another cost model and has to be looked at with figures over five years, not over one.
Where it behaves well.
Where there are many cameras and few people to watch them: a campus, a logistics centre, a hospital, an industrial estate. And where large areas have to be covered from few mountings, because its units with several sensors in one body cover a crossing of corridors or an open yard without spreading out four cameras with their four cables.
It pays off above all on the day something has to be found. Looking for a person in a light jacket in four hours of recording from sixty cameras is the work that really gets done with a video system, and it is exactly what this house has spent years polishing. And if the access control is theirs as well, cameras, doors and users end up in the same place, with a single record of who has looked at what.
Where we would not put it.
The commitment it brings is real, and there are three cases where it does not pay off.
As a stand-alone camera in an installation that is going to carry on with another house's video software. It is well made and it will work, but you pay for one half of the system and use the other: for that there are equivalent units that cost less.
At the difficult small points —an extreme backlight, reading a number plate at long distance, a scene with almost no light—, where what decides is the lens and the sensor and there are manufacturers more specialised in that. And in small installations, where building the whole system is out of proportion for eight cameras.
The analytics it carries inside, and where they end.
It is its own ground and it shows: classification of people and vehicles in the camera itself, an alert for intrusion into a zone, and search by appearance, which is what really changes the day to day of whoever reviews video. It has been running for years and it is well sorted.
The limit is in the kind of question that can be asked. It finds things that look alike and it raises alerts on rules that can be drawn over the image. It does not understand situations: that somebody is forcing a door, that a worker is working without the equipment today calls for, that what is out of place is the hour and not the spot.
What comes now is a recommendation with an interested party behind it, so it is declared: in the same house there is IRIS Neural, the NVMS from Infinity Neural, the other company in the group. It is our product. We would rather you knew it by reading this line.
What it brings is that understanding of the scene, and relating what happens on several cameras as a single story. They live together: this house's system governs the recording and the search, which is what the customer already has built and paid for. And if searching by appearance solves the problem, everything else is unnecessary and we say so.
What we do.
Here a good part of the work is making sure the final figure does not spring a surprise in year three.
- The complete sums before choosing: equipment, licences, real storage at the quality that is genuinely needed, and what gets paid each year. With the cloud part, the same calculation over five years.
- The choice between a unit with several sensors and several separate cameras, which is not a matter of taste: it depends on whether there is a point from which everything can be seen without a column or a parked lorry covering half of it.
- The lens and the field of view with the plan, and the calculation of the detail that reaches the exact place where somebody has to be recognised.
- The installation and the commissioning at the bad hour, into the sun and at night. And the tuning of the analytics on the real scene, with a week of alerts in view before anything is taken as good.
- The maintenance: firmware, passwords in the customer's name and an inspection of the field of view camera by camera on every visit.
The angle and the height are not argued about at the fitting.
With a unit of several sensors this is multiplied by four. Put in badly you do not lose one camera: you lose the four sensors at once, because they all look from the same point. A mounting chosen for the convenience of the cabling leaves two sensors looking at the roof of an industrial unit and another one facing a light fitting.
The angle, the height and the light decide for ever what will be possible with that image. It is calculated in the project, with the plan and the bad hours in front of you, because afterwards the correction is not moving a screw: it is paying for the labour again.
Where to go next.
Before comparing brands it is worth being clear about what you are going to ask of each image. That is what the area talks about.
Get started
Is what you need to search, or to see?
That is the question that makes this house interesting. Tell us how many cameras there are, how many people watch them and what happens on the day something has to be found in the recording. With that we will tell you whether the whole system pays off or only part of it.