AI Video Detection & GDPR
AI-based video detection captures traffic in greater detail than any other method – whilst at the same time raising questions about data protection. DataCollect’s ARGOS system resolves both issues: the artificial intelligence analyses traffic directly within the device (edge AI) and stores only anonymised object data. No actual video recording is required for this.
This is particularly crucial for local authorities and public sector clients: detailed, reliable traffic data with 100 per cent GDPR compliance.
How Edge AI video detection works
Unlike traditional camera systems, which transmit footage to a server for analysis, ARGOS processes the traffic scene directly within the device. This on-board analysis (Edge AI) detects, classifies and counts road users in real time – without any personally identifiable image data leaving the device.
- On-board processing: the AI runs on the device, not in the cloud
- Only anonymised object data is stored (class, direction, time, speed)
- An actual video recording is not necessary for the analysis
- Data is transmitted anonymously via LTE or Wi-Fi to myTrafficData 2
What ARGOS detects: 11 object classes
ARGOS’s AI distinguishes not only the size but also the type of road user. A total of 11 semantic object classes, including pedestrians and cyclists, enable a detailed analysis even of multimodal and complex traffic – at junctions, roundabouts and multi-lane main roads.
- Motorised traffic: cars, lorries, buses, motorbikes and other classes
- Active mobility: Pedestrians and cyclists
- Comprehensive detection of junctions, including turning patterns and source/destination traffic
- Real-time classification with no delay thanks to on-board processing
GDPR compliance in detail
ARGOS is 100% GDPR-compliant. The key point is that analysis takes place directly within the device; only anonymised object data is stored and transmitted. Personal image data – such as faces or number plates – is not transmitted to the cloud and is not required for traffic counting.
100 per cent GDPR-compliant: on-board analysis (edge AI), exclusively anonymised object data, no actual video recording required – no personal image data in the cloud.
- Data minimisation: only the object data required for counting is generated
- No personal image data in cloud processing
- Cloud platform myTrafficData 2 – Made and hosted in Germany
Why this matters for local authorities and government departments
Public sector clients are subject to particularly strict data protection requirements. A traffic counting system that collects personal video data quickly leads to the need for consultation with the data protection officer and to public reservations. ARGOS constructively circumvents this risk: as no personal image data is generated, its use is uncontroversial under data protection law – whilst providing comprehensive data for traffic planning.
Reliable data: Accuracy
Data protection must not come at the expense of data quality. ARGOS was tested in the Mobility Lab and achieved an accuracy of over 95 per cent – even under heavy traffic conditions with a daily traffic volume (DTV) of between 70,000 and 100,000 vehicles per day.
>95% accuracy in the Mobility Lab test – proven at a daily traffic volume (DTV) of 70,000 to 100,000 vehicles per day.
Recounting – can be adjusted retrospectively
If the requirements for an analysis change, the travel relationships can be adjusted in the cloud even after the measurement has been completed, and the data recalculated – retroactively for up to 90 days.
The recounting function is unique: counting zones are adjusted retrospectively and the results recalculated – without the need for a new measurement campaign.
Classification: ARGOS, SDR and DSD
DataCollect offers two measurement principles for traffic counting: the AI-based video detection system ARGOS and the radar-based SDR side-view radar. Both count and classify traffic, but differ in their detection geometry and depth. This should be distinguished from the DSD speed display: It primarily displays speeds and, from the DSD smart configuration level onwards, can optionally collect anonymised traffic data (volume and speed, including timestamps), but does not carry out any object classification.
| System | Function |
|---|---|
| ARGOS (AI video) | Area-wide counting and classification, 11 object classes, junctions |
| SDR side radar (radar) | Cross-sectional counting, speed and length classes |
| DSD speed display | Speed display (smiley); collects, from DSD smart onwards, optional anonymised traffic data (volume, speed, timestamps), but no object classification |
Frequently Asked Questions
Yes, ARGOS is 100% GDPR-compliant. The AI analysis takes place on-board directly within the device (edge AI); only anonymised object data is stored and transmitted. Personal image data is not transmitted to the cloud, and actual video recording is not necessary for traffic counting.
No. No actual video recording is necessary for traffic counting. ARGOS detects traffic via camera, but analyses the scene directly within the device and only stores anonymised object data such as vehicle class, direction, time and speed.
Edge AI means that artificial intelligence carries out the analysis directly within the device (‘on-board’), rather than sending image data to a server. This reduces data transmission, enables real-time results and forms the basis of GDPR compliance, as personal image data does not leave the device.
ARGOS distinguishes between 11 semantic object classes, including pedestrians and cyclists. This enables the differentiated detection of multimodal and complex traffic at junctions, roundabouts and multi-lane thoroughfares.
In the Mobility Lab test, ARGOS achieved an accuracy of over 95 per cent, verified under average daily traffic (ADT) conditions of 70,000 to 100,000 vehicles. The on-board analysis provides the data instantly.
The anonymised object data is stored on the myTrafficData 2 cloud platform, which is developed and hosted in Germany (Made and hosted in Germany). Upon request, the data can also be integrated directly into your own traffic management system.
Further information on comparing measurement principles: Traffic counting with radar or AI video. Product pages: ARGOS mobil and ARGOS permanent.
Would you like to clarify GDPR compliance for a specific measurement project?