If you are looking at Hikvision for people counting, you are looking at a camera-based method. Hikvision's public people-counting product pages describe dual-lens 3D stereo-vision cameras paired with deep-learning analytics to count visitors at entrances. That approach works, and it is well established, but it puts a camera at the door. For a growing number of buyers, the camera itself is the problem: it is what triggers CCTV signage, footage retention rules, a data protection impact assessment, and, in many European sites, a works-council conversation before a single count is produced.
This post is for buyers who want the count without the camera. It explains what a camera-free alternative actually is, why removing the image changes what you have to document, and how Ariadne counts without a lens at all. Comparisons cite public documentation; no client relationship or disparagement intended.
What is a camera-free alternative to Hikvision people counting?
A camera-free people-counting alternative records counts without capturing images of visitors. Ariadne does this with Hybrid Fusion: Time-of-Flight depth sensing at entries captures geometry rather than images, and patented phone-signal sensing follows movement inside, with both feeds fused in the platform. No MAC address is recorded by default, no device ID, no biometric data, and no camera is involved. Where a camera-based counter raises CCTV, retention, and works-council questions, a camera-free method changes what you have to document. Confirm any Hikvision people-counting detail against its own product pages before drawing a side-by-side, and verify accuracy on your own site rather than on a datasheet.
The rest of this post treats the camera as the deciding variable, because on the two systems it is the one difference that changes the paperwork, the privacy posture, and the internal approvals as much as the counting itself.
Why "camera-free" changes what you have to document
A people counter that sees is, in data-protection terms, a camera first and a counter second. Even where a vendor states that the counting analytics run on the camera itself and only a number is sent onward, the sensor is still an optical device pointed at people, and that is what most compliance frameworks react to. Removing the image removes a chain of obligations that a camera-based counter carries whether or not you ever look at the footage.
Four things change when there is no camera in the count:
- CCTV signage. A camera in a public or workplace space typically requires visible notice that recording is taking place, plus a documented purpose. A camera-free counter that captures no image does not create that footage in the first place, so there is nothing to sign about at that level.
- Retention. Video, even short-lived buffered video, forces a retention policy: how long frames are held, who can access them, and how they are deleted. A count that is derived from depth geometry and signal data, with no image ever formed, has no footage to retain.
- DPIA scope. A data protection impact assessment for a camera system has to weigh the risk of image capture and any potential for identification. A method that captures no image and no identifier narrows what the assessment has to cover. It does not remove the assessment, but it changes what is on the table.
- Works council. In many European organisations, installing cameras that observe staff or customers is a co-determination matter that a works council can slow or block. "It only outputs a number" is a harder case to make when the device is visibly a camera than when there is no camera present at all.
None of this makes a camera-based counter non-compliant. Plenty of organisations run them correctly. The point is narrower: the camera is the thing that opens the file. If your reason for shortlisting a Hikvision alternative is to shorten that file, the sensing method is where the difference lives, not the reporting dashboard on top of it. For the underlying distinction, see biometric vs non-biometric counting and camera-free people counting.
Hikvision people counting vs Ariadne, at a glance
Every Hikvision cell below reflects its public people-counting product documentation as published; confirm the current version on the vendor's own site before you rely on it. Ariadne's column is its canonical posture. Accuracy is deliberately left as a verify-on-your-own-site cell for both, because a datasheet number rarely survives a busy door.
| Axis | Hikvision people counting | Ariadne |
|---|---|---|
| Capture method | Dual-lens 3D stereo-vision camera plus deep-learning analytics (per public product pages) | Camera-free Hybrid Fusion: Time-of-Flight depth at entries plus patented phone-signal sensing inside, fused in the platform |
| Camera / images involved | Yes, an optical camera is the sensor | No camera; depth sensing captures geometry, not images |
| What it records about visitors | Verify on your own site against current product documentation | No MAC address by default, no device ID, no biometric data, no image |
| Documentation burden (CCTV / DPIA) | Camera-based: typically brings CCTV signage, footage retention, and a broader DPIA scope | No image captured, so no CCTV footage to sign about or retain; narrower DPIA scope |
| Verified-on-site accuracy | Verify on your own site | Verify on your own site |
The table exists to structure the questions, not to declare a winner. The one row that is genuinely different in kind, rather than degree, is the camera row: one method uses a lens, the other does not, and every downstream difference in the documentation column follows from that.
How Ariadne counts without a camera
Ariadne's method is worth stating precisely, because the whole camera-free case rests on it being true rather than a marketing gloss.
Ariadne measures this with Hybrid Fusion, its patented camera-free method. Time-of-Flight depth sensing counts every visitor at the entrances, capturing geometry rather than images, while patented phone signal sensing follows movement through the interior, detecting the signals a phone emits even in airplane mode, and tracks that movement to about one-metre precision. The sensor streams both feeds to Ariadne, where Hybrid Fusion combines them into one trajectory per visit and computes counts, dwell, and paths. The streams carry no identifier: no MAC address, no device ID, no biometric data, and no camera is involved. Identifiers are stored only when a visitor explicitly opts in, which keeps the method GDPR-friendly and outside biometric territory.
Two consequences matter for a buyer replacing a camera-based counter. First, there is no PII captured at the sensor. This is not anonymisation applied after the fact; there is no personal data collected in the first place that would need anonymising. Second, because no biometric identification or categorisation takes place, the method sits outside the biometric categories that the EU AI Act treats as high-risk under Annex III. For how counting maps to that regulation, see where counting sits under the EU AI Act, and for the wider shortlist, the systems comparison and Ariadne's camera-free counting.
How to switch from a camera-based counter
Switching sensing methods does not have to be a leap of faith, and it should not be. The honest way to compare a camera-based counter against a camera-free one is a like-for-like trial, not a datasheet argument.
Pick one representative entrance, ideally a busy one with the group traffic and lighting conditions that break weaker counters. Run the candidate system for a defined period, and over the same window collect a manual ground-truth count for a sample of hours: a person with a clicker, or a reviewed reference recording, counting the same door the system counts. Compare the system's number against the manual number for those hours, and you have accuracy measured on your traffic rather than on someone's test lab. Do this for both the incumbent-style approach and the camera-free candidate if you can, on the same door, and the comparison answers itself. To structure that properly, see run a trial first.
A migration also has practical pieces beyond accuracy: how historical data carries over into the new reporting, how the two systems' definitions of a "count" line up, and whether the install work at the door differs. Scope those in the pilot too, so the rollout after it holds no surprises.
FAQ
Do I need cameras to count people accurately?
No. Ariadne counts with Hybrid Fusion: Time-of-Flight depth sensing plus patented phone signal sensing, never cameras. Time-of-Flight captures geometry rather than images, and signal sensing captures no MAC address by default, so the measurement involves no video, no faces, and no biometric data.
Is a camera-free counter as accurate as a camera-based one?
Accuracy depends on the site, not the brochure, so verify it on your own doors. A depth-plus-signal method and a stereo-camera method both aim for high accuracy at the entrance; the way to know which holds up on your traffic is a like-for-like trial with a manual ground-truth count over the same hours.
What actually changes for compliance if I remove the camera?
The camera is what opens most of the file. Removing it means no CCTV footage to sign about or retain, and a narrower DPIA scope, because no image is ever captured. It does not remove your data-protection duties, but it removes the obligations that exist specifically because a camera is recording.
Does Ariadne capture any personal data at the sensor?
No personal data is captured at the sensor. The Time-of-Flight and signal streams carry no MAC address by default, no device ID, no biometric data, and no image. Identifiers are stored only when a visitor explicitly opts in, so there is no PII collected that would later need anonymising.
How do I compare a Hikvision people counter against Ariadne fairly?
Compare on four axes: what each captures about visitors, the documentation burden that follows from the sensing method, install model, and accuracy verified on your own site over a defined trial. Ground any Hikvision specification in its current public product pages, and hold both systems to the same manual ground-truth count on the same entrance.
---
