Public-Access Facilities (ERP) welcome thousands of visitors every day. Stadiums, museums, shopping centers, train stations, and concert halls all face a common challenge: How can they determine, in real time, how many people are in a given space and whether the regulatory capacity limit is being met?
GDPR-compliant video analytics provides an operational solution to this challenge. By analyzing video feeds from existing cameras, it measures occupancy without identifying individuals, without storing personally identifiable images, and without using facial recognition.
This guide details each step of the process: regulatory framework, technical principles, compliance criteria, on-site deployment, and performance metrics. XXII supports public access building managers in this process with its CORE platform, designed to make physical spaces as easy to understand as a dashboard.
A public access facility is a venue designed to accommodate the public: shopping centers, museums, performance venues, stadiums, train stations, or airports. Each public access facility is subject to a maximum occupancy limit, defined by fire safety regulations and safety commissions.
Measuring occupancy involves determining, at a given moment, the number of people present in a specific space or area. This data is essential for complying with occupancy limits, staffing appropriately, and anticipating peak visitor numbers.
Traditional methods (manual counting, turnstiles, infrared sensors) remain approximate or limited to a single entry point. Video analytics is a game-changer by analyzing the entire flow of people using cameras already installed on-site.
A large public access facility often has dozens, or even hundreds, of cameras distributed throughout its spaces. Video analytics leverages this existing infrastructure to simultaneously cover entrances, exits, circulation areas, and waiting areas.
This network makes it possible to determine occupancy rates zone by zone in real time, without adding dedicated sensors. This shifts the focus from sporadic counting to a spatial understanding of occupancy.
Video analytics detects silhouettes and movements. It does not recognize faces, does not store biometric data, and does not attempt to identify individuals. This non-biometric approach is a prerequisite for public-facing facilities committed to respecting the privacy of their visitors.
The result is an anonymized count, continuously updated, that feeds directly into operational management tools.
No new cameras are installed, and existing equipment is not replaced. The software layer connects to existing video streams via standard protocols (RTSP, ONVIF). This strategic choice accelerates the time between deployment and the first measurable results.
XXII designed its CORE platform specifically with this in mind: to leverage already-installed cameras to extract directly actionable operational data, without any additional hardware investment.
The GDPR governs all processing of personal data, including the analysis of video images. The CNIL distinguishes between several scenarios based on the degree to which individuals can be identified and the data retention period.
For an anonymized occupancy measurement system, the CNIL specifies that anonymization within a short timeframe (less than five minutes after collection) allows the processing to be based on the controller’s legitimate interest, without requiring individual consent. This is the most common scenario for occupancy measurement in public access buildings.
In 2022, the CNIL published its position on augmented cameras in public spaces. It reiterates that devices that do not process biometric data and that produce only aggregated and anonymous statistics benefit from a more flexible legal framework, according to the CNIL’s official position on augmented cameras.
This framework allows, for example, the real-time display of crowd levels to guide visitors toward less crowded areas. The objective is purely statistical and operational—no identification, no individual tracking.
An operator who deploys a video analytics-based occupancy measurement system must comply with several requirements:
This reflects the requirement for transparency that underpins XXII’s ethical and GDPR-compliant approach: making visible what is being measured and being explicit about what is not.
The analytics software connects to IP cameras already installed on-site via standard network protocols. No physical intervention on the cameras is required. The connection is software-based, allowing for deployment within a few days.
Computer vision algorithms detect human silhouettes within each camera’s field of view. They count entries and exits for each defined zone, distinguishing between entry and exit directions.
This detection relies on AI models trained to recognize shapes and movements, without using biometric features. No faces are analyzed, and no personal identifiers are extracted.
Raw data (source images) is processed in real time and is not retained. Only aggregated metrics are sent to the dashboard: number of people per zone, occupancy rate, and hourly trends.
This native anonymization is a design choice, not an afterthought. It is what allows us to balance operational performance with respect for personal data.
Operations managers access the data through a dedicated interface. They view the occupancy rate for each zone, receive alerts when a threshold is reached, and can review historical data to anticipate recurring peaks.
XXII’s CORE platform centralizes these metrics in a collaborative space accessible to both field teams and management. Sites are compared, action points are identified, and resources are adjusted.
The first criterion is regulatory compliance. Your solution must demonstrate that anonymization occurs in real time, that source images are not stored, and that no biometric data is processed. Request the vendor’s technical documentation and Data Protection Impact Assessment (DPIA).
Verify that the solution works with your current camera fleet, regardless of brand or model. A vendor that requires you to replace your video infrastructure will increase costs and prolong deployment times.
Counting accuracy is critical for monitoring occupancy levels. A deviation of just a few percent can skew a decision to close or open an area. Insist on accuracy data measured in real-world scenarios, not just in a laboratory setting.
A large-scale ERP system can involve hundreds of video streams. The solution must be capable of handling this volume without compromising latency or accuracy. This is a matter of software architecture, not just hardware power.
XXII processes hundreds of video streams simultaneously thanks to CORE’s scalable architecture, which can be deployed in SaaS, on-premises, or hybrid modes depending on your infrastructure constraints.
A good occupancy measurement tool is more than just a counter. It centralizes data from multiple sites, allows for comparisons between zones and time periods, and triggers customizable alerts (occupancy thresholds, zone saturation, excessive queues).
Facial recognition is not necessary to measure occupancy in a public access building. Counting algorithms detect silhouettes, not identities. This distinction is fundamental.
Using facial recognition in a public space would subject the operator to considerably stricter obligations (processing of sensitive data as defined by the GDPR, and the requirement for legislation authorizing the system). In practice, this type of data processing is prohibited in public buildings in the absence of a specific legislative framework.
This reflects a founding principle of the XXII approach: to develop tools for operational performance, not tools for identifying individuals. This distinction is not a regulatory detail; it is a principle that underpins all product decisions for the CORE platform.
Start by taking inventory of your cameras: number, model, protocol, and coverage. Identify critical areas (entrances, lobbies, stands, galleries) where occupancy measurement is a priority. This audit determines the scope of the initial deployment.
Each zone in your public access building has a maximum capacity, set by fire safety regulations or your own operational rules. Configure these thresholds in the solution to trigger automatic alerts as soon as a critical level is approached.
The video analytics software layer is deployed on your network infrastructure. It connects to the cameras, calibrates the counting zones, and adjusts detection parameters based on the field of view and foot traffic density.
With CORE, deployment takes just a few days, site by site, without disrupting the facility’s operations. Each project is built around measurable operational objectives from the very first weeks: workforce adjustments, capacity management, and optimization of foot traffic flows.
On-site teams (reception staff, security managers, site managers) must be trained to interpret the dashboard and manage alerts. Onboarding is quick, as the interface is designed for non-technical users.
The occupancy rate is the ratio of the number of people present to the maximum authorized capacity. It is the primary indicator for regulatory compliance and for determining the appropriate size of security teams.
Tracking entries and exits zone by zone helps identify bottlenecks, peak times, and distribution imbalances. This data informs decisions regarding the reallocation of human and material resources.
The average time spent in a zone (lobby, grandstand, gallery, exhibition space) reveals which areas are attractive and which are merely passed through. This KPI helps optimize layout, signage, and the placement of services.
The saturation rate indicates how often an area approaches or exceeds its critical threshold. Tracking this metric over time allows for anticipating seasonal patterns and adapting operational planning.
By centralizing these KPIs on a single interface, ERP managers gain a data-driven view of their spaces’ performance. XXII’s CORE platform aggregates this data by site and by network, making physical locations as measurable as a website.
The French market features several video analytics providers targeting ERP systems. Each takes a different approach to compliance, deployment, and technology models.
Some providers offer video analytics modules integrated into broader security management platforms. Others focus on people counting using dedicated sensors, which requires a hardware investment. Still others combine video analytics with IoT.
What sets XXII’s approach apart: a platform designed to run on existing cameras, without additional hardware, featuring native anonymization, non-biometric AI, and rapid deployment across large-scale, multi-camera sites. We have focused on four major sectors where real-time analysis of physical flows has a direct and measurable operational impact, including shopping centers, retail, transportation, and logistics.
A shopping mall connects its existing cameras to the video analytics platform. The dashboard displays, in real time, the number of visitors by zone, by floor, and for the entire site. Managers receive an alert when the capacity of a shopping arcade or lobby reaches a critical threshold.
This management approach reduces wait times at entrances, directs foot traffic to less crowded areas, and allows staff levels to be adjusted based on predicted foot traffic. Results are measurable within the first few weeks of deployment.
In a museum, measuring occupancy by zone allows for regulating access to the most popular galleries. Choosing the right video analytics partner ensures smooth visitor flow management while maintaining the quality of the visitor experience.
Curators and site managers can adjust tour schedules, staff placement, and signage based on real-time, hour-by-hour visitor data.
In a stadium, occupancy monitoring covers the stands, concession stands, walkways, and entry points. Real-time monitoring helps security teams quickly identify areas of excessive crowding and trigger the appropriate protocols. The same principle applies in train stations and airports, where the transportation sector benefits from real-time insights into passenger flow.
Analysis of historical data (attendance by event type, hourly peaks, bottleneck areas) helps improve security plans from one event to the next.
The European Artificial Intelligence Act (AI Act), which took effect in 2024, imposes new requirements on AI systems deployed in public spaces. Video analytics systems used for anonymized counting benefit from a less stringent regulatory framework compared to biometric recognition systems, which are classified as high-risk.
Preparing for compliance with the AI Act involves documenting the purposes of data processing, ensuring algorithmic transparency, and maintaining human oversight of operational decisions. A vendor that incorporates these requirements into the design of its platform from the outset makes it easier for you to achieve compliance, in accordance with the principles of the European General Data Protection Regulation (GDPR).
XXII incorporates these principles into the architecture of CORE: explainability of results, traceability of processing, and the absence of biometric data. This rigorous approach prepares retail stores for future requirements without operational disruption.
Occupancy measurement via video analytics provides shopping mall managers with a precise management tool that complies with the GDPR and can be deployed on existing infrastructure. It is based on a clear principle: analyzing foot traffic, not individuals.
The choice of a software provider hinges on verifiable criteria: native anonymization, compatibility with your cameras, counting accuracy, the ability to handle high traffic volumes, and documented compliance. These criteria distinguish an operational solution from a simple counting tool.
XXII is driven by the conviction that high-traffic locations have a lot to say, and that the right technology must simply be able to interpret them—at the right time, for the right reasons. The CORE platform translates this conviction into measurable results for retail, deployment after deployment. You can also learn how logistics benefits from the same approach.
Yes, provided that the data is anonymized promptly and no biometric data is processed. The CNIL authorizes this type of system on the basis of legitimate interest when anonymization occurs within a few minutes.
XXII implements native anonymization within its CORE platform, which means that source images are never stored and no faces are analyzed.
No. Video analytics connects to the IP cameras already installed at your site via standard network protocols. XXII deploys CORE directly on your existing infrastructure, which speeds up deployment and avoids additional equipment costs.
Video analytics detects silhouettes, movements, and the flow of people. It does not seek to identify who is present. Facial recognition, on the other hand, processes biometric data to identify individuals, which is subject to strict restrictions in public gathering places.
The key metrics include real-time occupancy rate, entry and exit flows by zone, average dwell time, and alerts when capacity limits are exceeded. XXII centralizes these KPIs in the CORE interface for site-by-site operational management.
Deployment depends on the size of the site and the number of cameras to be connected. With XXII’s CORE platform, software installation takes just a few days, without disrupting the facility’s operations.