WestPoint · AxxonSoft
AxxonSoft makes video software with a different idea from almost everybody else's: the analysis runs on the server, over the cameras you already have, old or new. That opens doors other platforms close and brings its own invoices — in machines, in tuning and in questions from purchasing.
That is the scene where this platform clicks straight away.
There are large installations —a factory, a municipal complex, a port, a hospital— with a camera estate built up in layers over fifteen years: different brands and generations, none of them with analytics inside. Asking for analytics there usually means changing the cameras, and changing two hundred cameras is a project that does not get approved.
AxxonSoft goes at that: the detectors run on the server, so what is needed in order to analyse is not a new camera, it is computing capacity on the machine. A camera from ten years ago that still gives a decent image can take part in a count or in a zone detection.
And it brings two tools that you recognise the moment you use them once: one lets you review hours of recording in minutes, because it puts what happened at different moments into the same view; the other searches inside the archive by what happened and not by the time —an object that entered this zone, a vehicle of this colour—. Anyone who has reviewed a whole weekend of recording knows what that is worth.
It also has a second platform, older and of another nature, for bringing video, access, fire and plant systems together in one console. It is the route for industrial or critical installations where the video is one part of the whole.
A classic platform, on your servers, with one particularity that conditions everything: the analytics eat machine.
Two conversations that cannot be left until last. Both are about purchasing and compliance.
The first is about origin and supply chain. In many European specifications —and in almost all of those for the public sector and critical infrastructure— the question today is where the software comes from, which company answers for the contract, where the support is and what guarantees there are of continuity and of updates. With AxxonSoft that question is asked before designing and answered with documentation from the manufacturer and from its representation in your country. We do not answer it: we make sure it has been answered before anybody signs.
The second is about the security of the system itself, and it applies to any platform with a lot of old estate behind it. A video system is software exposed on a company's network: you have to know how often the manufacturer publishes fixes, who is going to apply them, which network the servers and the cameras live on and who gets in from outside. When the answer is «the installer looks after that», there is no answer.
And one point in its favour: the video stays at home. There is no compulsory cloud and no subscription governing the system, and for an installation that has to explain where the images are, that simplifies the file. Remote access is solved with a VPN and with the network properly separated, which is IT work — and it has to be said that it is.
When the analytics run on the server, the server stops being a place where files are kept. Every detector on every camera consumes, and the ones that work with neural networks ask for a graphics card. The sum is not «how many cameras»: it is how many cameras, with how many detectors each, at what resolution and at how many images per second.
That is why the budget has two parts and the second gets forgotten: licences and machines. We have seen installations where the analytics were bought and never fully switched on, because the server was enough to record and not enough to think. The advantage of the same model is real: analytics are added without touching a camera or climbing a pole, so counting vehicles at an entrance that is already recorded is licence and capacity, not building work.
It is a platform for a large installation with an inherited estate. Outside that, there are more convenient options.
You have many cameras of different brands and ages and you are not going to change them. If you need analytics over what is already in, or to review the archive often and quickly. If the video has to live in the same console as access, fire or plant systems. If you want everything to stay on your servers. And if there is a technical team —yours or ours— able to hold up a platform with a lot of tuning.
You are looking to bring twenty small premises into service with little work: this is the opposite. Nor if the purchasing policy demands a manufacturer of a particular origin, or a long list of local references and certifications: that conversation has to be won beforehand and sometimes it is not won. And nor if nobody is going to tune the detectors, because server-side analytics left untuned produce alerts that end up being switched off.
Here the analytics are not an add-on, they are the argument: a wide catalogue of detectors on the server —zones, crossings, loitering, object left or removed, people counting, occupancy, vehicles and number plates, smoke and fire in some versions— and forensic search, which consists of asking the archive about what happened and not about the time.
What can be added: the manufacturer's own by modules, the analytics of whatever good cameras there are and whatever is integrated through the programming interface. And every thing that is added comes back to the same place: machine is needed.
The ceiling, which with a catalogue this size has to be said more plainly: these are detectors. They detect defined patterns and they need somebody to tune them for that camera, that light and that scene. They work very well on the well-bounded —a zone, a line, a count— and they fall short when the question is interpretative: whether what is going on there is normal or not, with three things at once. No number of detectors reaches that, and promising it is what gets the alerts switched off a month later.
And the declaration we owe you: in the same house there is IRIS Neural, the NVMS from Infinity Neural —the other company in the group, our own product, and we say so before recommending anything—. On many installations they live together: the VMS governs cameras, recording and retention, and IRIS brings the understanding of the scene, which is the interpretative question we have just talked about. With an open platform on your own servers the coexistence is technical and without surprises: both parts look at the same streams, without duplicating cameras and without taking video out of the installation.
On this platform the integrator shows more than on any other. For better and for worse.
It is the platform that demands most of whoever maintains it, and the one that gives most back when there is somebody behind it.
Almost nobody chooses video software by looking at a single platform. These are neighbours in the same conversation, and the index leads to all twenty.
Get started
If the answer is a lot and from several generations, this is one of the conversations worth having. Tell us what is installed, what you would like to be able to detect and what your purchasing department demands, and we will tell you what can be done with what you have, what machine would be needed and where the limits are.
