Video Analytics Sourcing: False Alarms, Licensing, AI Act

Video Analytics Sourcing: False Alarms, Licensing, AI Act

By Zoey Zhang, Founder & Lead Sourcing Specialist

Sourcing Guide
Technical China Sourcing Vlog: 2-Day BOQ & QC for CCTV, Barrier Gates, Smart Locks & IP Phones

    One camera can carry 4 MP, starlight and “AI support” on the same specification line — but whether it merely raises alarms, or raises alarms somebody can act on, is a whole set of procurement actions apart.

    The three regrets we see most: the analytics rules needed a separate licence that was not on the quotation; three months after handover the guard turned the alarm sound off; the project finished and the owner’s legal team asked whether the facial recognition was compliant. This article settles three things only: where the analytics runs, how false alarms get verified, and where the compliance red lines are. Camera selection — resolution, lens and DORI pixel density — is covered in work out the resolution before you buy; this one does not repeat it.

    CCTV video analytics alarms being reviewed by an operator on the monitoring wall

    1. Analytics runs in three places, and the quote can differ three-fold

    The same “line crossing alarm” can run inside the camera, inside the recorder, or inside a VMS on a server. They are not upgrade steps of one product; they are three different businesses:

    DimensionCamera side (edge)Recorder side (NVR)Server side (VMS)
    ComputeFixed by the chip; usually 1–2 rules per channelMedium, limited by channel countScales with CPU/GPU, cards can be added
    When the network is downStill alarms and records locallyWorks inside the LANDepends on the server being online
    Cost per channelBuried in the camera price, invisibleBuried in NVR channel countLicensed per channel, visible every year
    Changing a ruleLog in to each unit; 50 channels is a day’s workBulk changes, still limited by modelOne push covers the site
    FitsUp to 20 channels, single rule typeUp to 50 channels, central managementAbove 50 channels, counting and search

    Settle this column first and every price afterwards becomes comparable. Video analytics is a subsystem of the security package — how it splits and how responsibility is drawn across the rest is set out in what an ELV package covers.

    2. “AI features” are not a spec line, they are a licence

    The camera brochure writes AI as a parameter; the supplier quotation writes it as something else entirely — a licence. Three forms, and they are not interchangeable:

    1. Free with the unit: pixel-level decisions such as motion detection and privacy masking. No model involved, normally not charged, and equally unable to filter leaves and rain.
    2. Licensed per channel: the mainstream form of server-side analytics, priced as “camera channel × year”, so 20 extra channels means 20 extra licences.
    3. Licensed per function module: face, plate and people counting sold separately, some tiered by analysed frames per second.

    Seven items the quotation has to answer:

    AskWhy it matters
    Is licensing per channel, per device or per functionDecides the cost of expanding three years from now
    Term: perpetual / annual / subscriptionA missed annual licence is a fixed cost every year
    Does the licence transfer when a camera is replacedA brand change normally means buying again
    Are software version upgrades charged separatelySome charge per major version
    Maximum number of rulesOne rule per zone, or eight
    Does it include tuning and on-site commissioningInstalled but never tuned means the factory default false alarm rate
    Must the analytics and the camera be the same brandOnce bound, there is no negotiating room left

    3. Break “95% accuracy” into two numbers

    A supplier says 95% accuracy. You cannot accept against it, because it never says what the denominator is. Split it:

    • Precision = true events ÷ (true events + false alarms): how many of the alarms were real.
    • Recall = true events ÷ (true events + misses): how much of what actually happened was caught.

    Take an estate perimeter of 8 zones, each averaging 2.5 false alarms per night:

    ItemNumber
    False alarms per night8 × 2.5 = 20
    Per month (30 days)600
    Per working month (22 days)440
    3 minutes of guard review each30 hours a month ≈ 3.8 eight-hour shifts

    In other words, close to four full shifts a month are spent walking over to look and walking back. Run the same 20 alarms a day at 90% precision and the wasted trips fall to 2 a day, 60 a month — the difference is not the camera’s ability, it is how the parameters were tuned. So the contract cannot say “95% accuracy”; it has to say “under the scenes and lighting conditions in the annex to this agreement, precision ≥ X% and recall ≥ Y%, judged by the record sheet in the annex”.

    4. How to run a 30-day POC

    The purpose of a POC is not to prove the system works; it is to produce numbers you can put in the contract.

    StageWhat happensDeliverable
    Week 1Agree the scene list: lighting (backlight, night IR switch-over), weather (rain, fog), target (person, vehicle, left object), zone (walkway, boundary, racking)Scene table, at least 3 points per scene
    Week 2Install, run one week on default parameters, no tuningRaw alarm log
    Week 3Tune per scene (sensitivity, minimum object pixels, time threshold), run againTwo log versions, before and after tuning
    Week 4Review: play back every alarm and mark true or false; compute precision and recallPOC report plus video evidence

    The record sheet needs at least these columns: time, zone, alarm type, true event yes/no, reviewer, notes — and the review must be done on playback, not on the guard’s impression. Image quality on the camera side decides the analytics outright: backlight, obstruction and a dirty lens cannot be recovered by any algorithm. The checks at arrival and installation are in the surveillance camera QC process, and full-unit vetting and burn-in in PTZ camera factory audit.

    Dome camera mounted on an external wall — edge analytics depends on picture quality

    5. Whether the metadata moves decides how locked in you are

    The output of analytics is not video, it is metadata: which frame, which object, what class, which coordinates. Three questions decide whether it crosses brands:

    ONVIF ProfileWhat it coversRelation to video analytics
    SBasic video streamingPicture only, no analytics results
    GRecording and searchRetrieving recorded video
    TAdvanced streaming (including H.265 and metadata)Carries some metadata
    MAnalytics, metadata and events (released June 2021)Standardised analytics configuration, object classification, plate/face/human/geo metadata, events over MQTT

    M is the only profile that standardises analytics results enough to cross manufacturers: change camera brand and the VMS integration does not need rewriting, and events can drop straight into an IoT platform for automation. If the supplier only offers a private HTTP API plus a private SDK, your analytics results are nailed to that system — changing brand means starting again. The procurement action: write into the technical requirements that “both cameras and VMS shall declare ONVIF Profile M support, with the conformance statement in the delivery list”, and do not accept a vague “we are ONVIF compatible”.

    6. Bandwidth and storage: do not size only the main stream

    Analytics needs another stream of picture, and the metadata travels over the network too. A 50-channel site, worked through:

    ItemCalculationResult
    Second stream for analytics512 kbps × 5026 Mbps uplink
    Metadata stream1–20 kbps × 500.05–1.0 Mbps (small, but keep it on its own VLAN)
    Main stream4 Mbps × 50200 Mbps
    Storage per channel per 24h4 Mbps × 86400 s ÷ 842 GB
    50 channels per day42 GB × 502.1 TB

    The metadata volume is small, but it is what makes “search by person, search by vehicle” possible — recorded video search runs on it. What actually eats the budget is the main stream; how retention days and bitrate are set is in how to size CCTV storage for a 500-room hotel. Confirm the uplink separately: plenty of projects have gigabit switch ports but a single 100 Mbps uplink, and as soon as the analytics streams are added the picture breaks up.

    Switches and patch leads — the second analytics stream needs its own uplink calculation

    7. Compliance: three red lines belong on the purchase list

    The prohibitions in Article 5 of the EU Artificial Intelligence Act (Regulation (EU) 2024/1689) have applied since 2 February 2025. Four of them bear directly on security cameras:

    ClauseWhat is prohibitedWhat to do in procurement
    5(1)(e)Scraping surveillance or internet imagery to build or expand face databasesRefuse factory-preinstalled face databases / image search
    5(1)(f)Emotion recognition in workplaces and education (except medical and safety uses)Strike out emotion or psychological state analysis
    5(1)(g)Inferring sensitive attributes such as race, politics, religion or sexual orientation from biometricsStrike out gender/age and crowd-attribute statistics
    5(1)(h)Real-time remote biometric identification in public spaces (law enforcement only, very narrow exceptions)Face recognition only on controlled access lanes, never in public areas

    On the data side there is one more that gets missed: GDPR Article 35(3)(c) requires a data protection impact assessment (DPIA) for large-scale systematic monitoring of publicly accessible areas. The duty sits with the owner / data controller, but if the buyer cannot hand over a statement of what the cameras capture, how long it is kept and who can see it, the DPIA stalls at the supplier.

    The cheapest route is to put the feature list on the first page of the technical requirements: buy only the analytics you need (line crossing, intrusion, left object, people counting), and never preinstall face or attribute functions. Chinese firmware often ships with them switched on; disabling them before export with a written record is cheaper than explaining afterwards.

    8. Eight questions that must be closed with the supplier

    1. Does the analytics run in the camera, the recorder or the server? What is the rule limit per channel?
    2. Is licensing per channel, per device or per function? For how long? Does it transfer on replacement?
    3. Which ONVIF Profiles are supported, and can you supply a Profile M conformance statement?
    4. What is the false alarm behaviour on default parameters, and do you have measured records from a comparable scene?
    5. Does the offer include on-site tuning and 30-day POC support? Who does the tuning?
    6. What bitrate and resolution does the analytics second stream use, and how is the uplink sized?
    7. Can face, emotion and attribute functions be disabled at the factory with a written statement?
    8. Are software upgrades and rule template updates charged separately?

    9. Six traps that show up most often

    TrapConsequenceHow to avoid it
    Reading “AI support” and never asking the rule count50 channels able to run one ruleWrite the rule limit into the quotation
    Licensing quoted for one year onlyA budget top-up in year twoContract the three-year total cost of ownership
    Accepting on factory default parametersNobody owns the false alarm ratePut the POC record sheet in the annex
    Ignoring the second streamUplink saturated, picture stutteringSize the uplink separately
    Camera and VMS forced to one brandNo negotiating room leftRequire Profile M or an open interface
    Face and attribute functions left on by defaultAn EU project that cannot pass complianceFeature list up front, written confirmation of disablement

    Video analytics is not the purchase of a “smarter camera”; it is the purchase of a capability that can be accepted, expanded and explained. Put the scenes, the rules, the licensing and the record method into the enquiry document and every quotation lands on the same basis. If you already have drawings or a point schedule, send the scenes you want analysed — which zones, which targets, which hours — and we will come back with a comparable configuration and cost structure by point count and rule count, within 48 hours. For how this lands in a 500-room hotel project, see the delivery record for this project.

    Frequently asked questions

    Both the camera and the recorder claim "AI support" — which side should run the analytics?

    Look at rule count and the expansion plan. Camera-side analytics uses no server and still raises local alarms with the network down, but the compute is fixed by the chip — usually one or two rules per channel. Recorder or server-side analytics scales by channel and rules can be changed in bulk, but the machine has to stay online. Under 20 channels with only line crossing and intrusion, camera side is enough; above 50 channels, or if you need people counting or plate recognition, put the analytics on the server and leave the camera to deliver a clean picture.

    A supplier says 95% accuracy — can that number go into the contract?

    No, because it does not say 95% of what. 95% precision means one wasted alarm in twenty; 95% recall means five real events missed in a hundred — on the same system those two numbers are often far apart, since conservative tuning cuts false alarms and raises misses. The contract should state "under the scenes and lighting conditions in the annex, precision not below X% and recall not below Y%", with the POC record sheet attached.

    Our project is in Southeast Asia — does the EU AI Act apply to us?

    If the equipment ends up on the EU market, or the owner is an EU entity, it does. The prohibitions in Article 5 have applied since 2 February 2025 — many Chinese cameras ship with marketing features such as face database building, emotion recognition and gender/age attributes, and those are high-risk or outright prohibited uses in the EU. Crossing them off the feature list at purchase, or getting a written statement that they are disabled, is far cheaper than explaining afterwards.

    How long does a POC have to run to count?

    30 days, spanning clear weather, rain and at least one change of night scene. Anything under two weeks only proves the system can raise an alarm, not what its false alarm rate is — most false alarms come from sudden lighting change, rain and snow, moving branches and the night IR switch-over, and the frequency of those events needs time to settle. Aim for at least 50 real events recorded in the 30 days, otherwise the confidence in precision and recall is too low to hold in a negotiation.

    Can the analytics survive if we change camera brand?

    Usually not. Camera-side analytics is burned into the firmware and goes with the unit; server-side VMS analytics licences are normally tied to camera channels or device serial numbers, so a brand change means buying channel licences again. Asking "if we replace 20% of the cameras within three years, how is licensing charged" before signing saves a budget top-up later.

    Next step

    Need a quote or a technical answer?

    Tell us the product and the quantity. Our team in Guangzhou sends back pricing, MOQ and lead time within one business day.

    Guangzhou-based sourcing team · Replies in English within 1 business day

    Share this guide:
    WhatsApp