Most analytics platforms still ride on cameras. Ariadne's Hybrid Fusion pairs anonymous signal sensing with Time-of-Flight depth. Same insights as a camera build, none of the privacy cost or the lens-by-lens install.
Looking for the product overview? Start on the people counting page. This page is the deeper comparison for buyers already weighing sensor approaches.
The alternative
Camera and LiDAR
Most competing platforms
With Ariadne
Ariadne
Hybrid Fusion: signal sensing and Time-of-Flight
Accuracy
85-95% in good conditions. Drops fast under lighting changes, occlusion, and dense crowds.
Accuracy
Up to 99% in optimal conditions. 95%+ in typical retail and airport deployments. Holds up in low light and at peak density.
Privacy
Captures faces, bodies, and movement patterns.
Privacy
Anonymous device signals only. No PII ever stored.
Hardware
Cameras at every viewing angle. Cabling, mounting, lens upkeep.
Hardware
Compact signal sensors plus ToF units. Fewer pieces, smaller footprint, zero cameras.
Coverage
Only what each lens sees. Blind spots between cameras.
Coverage
Signal sensing reaches across whole venues. ToF validates counts at choke points.
Conditions
Lighting-dependent. Occlusion drops accuracy.
Conditions
Signal works in low light and crowded scenes. ToF stays accurate in any condition.
Compliance
GDPR / CCPA / EU AI Act friction. Ongoing legal risk.
Compliance
GDPR by design. TÜV-certified. EU AI Act ready.
Sensor technology comparison · 2026
Every sensing technology, side by side
A visitor counting system is only valuable if it is accurate, privacy-safe, and actionable. Most technologies force a trade-off: accurate door counts (3D sensors) or journey analytics (signal and Wi-Fi methods) or operational simplicity (beams). Here is how each one stacks up.
Hybrid Fusion (Ariadne)
Recommended
Up to 99% (typically 95%+)
Privacy
Privacy-first (Very High)
GDPR safe
Install
Fast
Best for
Airports, large retail, malls, smart cities, complex layouts
Pros: Combines both: ToF validates true counts at choke points while patented signal sensing adds continuity for multi-zone journeys. Measures footfall, occupancy, dwell, flow, queues, and supports staff exclusion.
Cons: Needs a short calibration window per site for the signal layer to settle.
Time-of-Flight (ToF) Depth Sensing
98 to 99.5% (entrance counting)
Privacy
High (depth data, not RGB video)
GDPR safe
Install
Fast
Best for
Standard entrances, doors, corridors
Pros: Strong privacy-by-design option for counting at entrances. Depth sensing works in low light and avoids capturing identifiable images.
Cons: Primarily line-crossing and entrance metrics. Limited journey analytics across multiple zones without an additional tracking layer.
3D Stereoscopic Video (Active Stereo)
70 to 99%+ (deployment dependent)
Privacy
Medium (camera-based)
Install
Complicated
Best for
High ceilings, busy entrances
Pros: Stereo depth reduces false counts from shadows and improves separation versus monocular video in complex scenes.
Cons: Still camera-based, which increases privacy and compliance burden and often requires more tuning and infrastructure than ToF.
mmWave Radar (Presence / Occupancy)
Great for occupancy; moderate for headcounts
Privacy
Very High (no images)
GDPR safe
Install
Fast
Best for
Washrooms, meeting rooms, HVAC occupancy
Pros: Excellent presence detection (even micro-movements) and works in darkness. Strong choice when the KPI is 'is anyone here?'
Cons: Harder to separate individuals in dense crowds. Typically not the best fit for decision-grade entrance counting.
Monocular AI Video (2D)
85 to 95% (scene dependent)
Privacy
Low to Medium (video processing)
Install
Slow
Best for
Security plus basic visitor counting
Pros: Can reuse existing CCTV and add visual verification or classification.
Cons: Accuracy is sensitive to lighting, occlusion, and camera angle. Capturing identifiable people on video increases GDPR obligations and governance overhead.
Fisheye Cameras (360 degree)
Trend-level (varies widely)
Privacy
Low to Medium (video)
Install
Slow
Best for
Open floors, large spaces (with specialist tuning)
Pros: Wide field-of-view with overhead mounting can reduce occlusions and support heatmaps.
Cons: Radial distortion makes detection and tracking harder. High accuracy often requires distortion-aware models, careful calibration, and sometimes multiple cameras.
Wi-Fi / Bluetooth Device Counting
Sample-based (not absolute footfall)
Privacy
Low to Medium (device identifiers)
Install
Fast
Best for
Recurrence and dwell (in compliant setups)
Pros: Can estimate recurrence and dwell over larger areas than a single doorway.
Cons: Counts devices, not people (phones may be absent or off). Wi-Fi tracking often involves personal data under GDPR and needs careful legal plus technical controls.
Active Infrared (Break-Beam)
80 to 95% (depends on doorway and flow)
Privacy
High
GDPR safe
Install
Fast
Best for
Single door, low-traffic entrances
Pros: Simple and low-cost way to do basic counting at a narrow entrance.
Cons: Side-by-side or group entries collapse into a single count. Limited analytics beyond basic in and out totals.
Technology
Accuracy
Privacy
Install
Best for
Pros and cons
Hybrid Fusion (Ariadne)
Recommended
Up to 99% (typically 95%+)
Privacy-first (Very High)
GDPR safe
Fast
Airports, large retail, malls, smart cities, complex layouts
Pros: Combines both: ToF validates true counts at choke points while patented signal sensing adds continuity for multi-zone journeys. Measures footfall, occupancy, dwell, flow, queues, and supports staff exclusion.
Cons: Needs a short calibration window per site for the signal layer to settle.
Time-of-Flight (ToF) Depth Sensing
98 to 99.5% (entrance counting)
High (depth data, not RGB video)
GDPR safe
Fast
Standard entrances, doors, corridors
Pros: Strong privacy-by-design option for counting at entrances. Depth sensing works in low light and avoids capturing identifiable images.
Cons: Primarily line-crossing and entrance metrics. Limited journey analytics across multiple zones without an additional tracking layer.
3D Stereoscopic Video (Active Stereo)
70 to 99%+ (deployment dependent)
Medium (camera-based)
Complicated
High ceilings, busy entrances
Pros: Stereo depth reduces false counts from shadows and improves separation versus monocular video in complex scenes.
Cons: Still camera-based, which increases privacy and compliance burden and often requires more tuning and infrastructure than ToF.
mmWave Radar (Presence / Occupancy)
Great for occupancy; moderate for headcounts
Very High (no images)
GDPR safe
Fast
Washrooms, meeting rooms, HVAC occupancy
Pros: Excellent presence detection (even micro-movements) and works in darkness. Strong choice when the KPI is 'is anyone here?'
Cons: Harder to separate individuals in dense crowds. Typically not the best fit for decision-grade entrance counting.
Monocular AI Video (2D)
85 to 95% (scene dependent)
Low to Medium (video processing)
Slow
Security plus basic visitor counting
Pros: Can reuse existing CCTV and add visual verification or classification.
Cons: Accuracy is sensitive to lighting, occlusion, and camera angle. Capturing identifiable people on video increases GDPR obligations and governance overhead.
Fisheye Cameras (360 degree)
Trend-level (varies widely)
Low to Medium (video)
Slow
Open floors, large spaces (with specialist tuning)
Pros: Wide field-of-view with overhead mounting can reduce occlusions and support heatmaps.
Cons: Radial distortion makes detection and tracking harder. High accuracy often requires distortion-aware models, careful calibration, and sometimes multiple cameras.
Wi-Fi / Bluetooth Device Counting
Sample-based (not absolute footfall)
Low to Medium (device identifiers)
Fast
Recurrence and dwell (in compliant setups)
Pros: Can estimate recurrence and dwell over larger areas than a single doorway.
Cons: Counts devices, not people (phones may be absent or off). Wi-Fi tracking often involves personal data under GDPR and needs careful legal plus technical controls.
Active Infrared (Break-Beam)
80 to 95% (depends on doorway and flow)
High
GDPR safe
Fast
Single door, low-traffic entrances
Pros: Simple and low-cost way to do basic counting at a narrow entrance.
Cons: Side-by-side or group entries collapse into a single count. Limited analytics beyond basic in and out totals.
Three short notes on where each sensing approach earns its place, and where it does not.
Why Hybrid Fusion is built for decision-grade counting
Most sensing technologies excel in one area and fall short in another. ToF-only sensors can be highly accurate at entrances, but they do not connect a visitor's journey across multiple zones. Signal and Wi-Fi methods can show movement patterns, but they are often sample-based and add privacy and compliance complexity. Ariadne's Hybrid Fusion solves the trade-off by combining patented signal sensing for continuity across zones and floors with Time-of-Flight to validate the true headcount at key choke points.
Privacy-first counting: depth events vs video vs Wi-Fi tracking
Camera-based counting can be effective, but video of identifiable people is personal data and increases GDPR obligations and operational overhead. Wi-Fi tracking can also process personal data (including location and trajectory data) depending on how it is implemented. Depth-based ToF counting reduces privacy risk by relying on depth information rather than RGB video. Ariadne goes further by avoiding stored video and applying real-time de-identification at the point of detection, producing aggregated metrics designed for privacy-first deployments.
Where mmWave radar fits, and where it does not
mmWave radar is excellent for presence and occupancy because it detects very fine motion, even when people sit still. But for accurate entrance counting in dense foot traffic, optical depth sensors typically provide clearer separation and validation. Use mmWave when the KPI is room occupancy (washrooms, meeting rooms, energy automation), and use ToF or Hybrid Fusion when you need decision-grade counts plus operational KPIs like dwell, flow, and queues.
See Hybrid Fusion against your venue
Bring your floor plan and your current numbers. We will map out where ToF, signal sensing, or both belong, and what accuracy you should expect.