AI Video Analytics for Retail

Smarter Retail Operations.
From Customer Flow to
Parking Management.

Retail environments generate valuable information every day — customer movement inside the store, vehicles in the car park, queues at the checkout. Much of it is never measured consistently. Metrici turns IP camera feeds into operational data you can act on: count visitors, see how they move, track queues and occupancy, manage parking spaces and automate vehicle access.

Request a retail demo Explore retail applications
People counting & movement analysis Occupancy and queue alerts Parking-space detection LPR vehicle access Works with compatible existing cameras Runs on your own infrastructure

Built for shopping malls and retail parks, supermarkets and stores, and retail chains that need consistent reporting across locations.

ZONE 1 · HIGH ZONE 2 COUNTING LINE A ENTRANCE SHOPFLOOR ENTRANCE A IN TODAY 1,284 OUT TODAY 1,197 ON FLOOR NOW 87 ZONE LOAD Z1 Z2 Z3 Z4 Footfall & zone activity In / out counting · dwell time · density by zone · historical patterns
CHECKOUT AREA · CAMERA 7 LANE 1 2 UNDER LIMIT LANE 2 3 UNDER LIMIT LANE 3 6 THRESHOLD 5 QUEUE ALERT Lane 3 · 6 customers Threshold of 5 reached Sent to duty manager STORE OCCUPANCY FRONT OF STORE FASHION GALLERY LIMIT 320 · NOW 268 Queue & occupancy monitoring Configured thresholds · real-time alerts · staff and lane decisions
A1 A2 A3 A4 A5 B1 B2 B3 · RM B4 B5 PPD · DECK 1 ENTRY LED DISPLAY 42 free · Level 1 Level 2 → 118 free DECK LOAD 318 OF 360 OCCUPIED ENTRY 10:41 · B4 DWELL 00:52 RESERVED B3 · PLATE CHECK Parking-space occupancy Parking Place Detector per space · availability guidance · reserved bays
DELIVERY GATE · LOADING AREA ACCESS B 12 PDF SUPPLIER PLATE RECOGNITION B 12 PDF WHITELIST MATCH Rule · supplier, 08:00-12:00 Barrier released · 0.23 s AUTOMATED ACTIONS BarrierOPEN LED totemDOCK 3 System eventJSON Vehicle access & deliveries LPR recognition · whitelist rules · barriers, displays and business systems

Illustrative overlays. Panels show the kind of information each application reports.

02

Retail decisions are made on incomplete information

A retail operator manages footfall, staffing, service levels, parking capacity and vehicle movements at the same site — usually with data sources that were never designed to work together. The result is decisions taken from estimates, sampled counts and memory of last Saturday.

Visitor flow

You do not know how many people actually enter

Door counters and manual samples give rough totals for one entrance on one day. There is no consistent record of how visitors move between zones, which areas they pass without stopping, or how long they stay. Marketing and leasing decisions end up based on assumptions rather than measurement.

Service level

Queues are noticed only when someone complains

Checkout wait times are usually discovered by customers, or by a supervisor walking the floor. By then the transaction is lost. Without a measured queue length per lane, staffing adjustments are reactive and cannot be compared across shifts, days or locations.

Space planning

Zone performance is judged by impression

Which corridor carries traffic, which gallery is being bypassed, which layout change actually improved activity — these questions are difficult to answer retrospectively when no zone-level activity history was recorded. Layout and tenant-mix decisions are therefore hard to justify.

Car park

Drivers circle while spaces sit free

In a large multi-level car park, occupancy is often known only as a barrier counter — a total, not a per-level or per-zone figure. Drivers waste time searching, congest internal lanes, and guidance displays either do not exist or show numbers nobody trusts.

Access control

Deliveries and tenants are handled manually

Supplier and tenant vehicles are admitted by guards checking lists, badges or phone calls. Each access event is a manual decision, records are paper or spreadsheet based, and misuse of reserved or tenant-only areas is discovered late, if at all.

Existing cameras

Cameras record, but nobody analyses

A retail site typically already runs dozens of IP cameras for security. Their output is reviewed after an incident, never turned into operational data. The infrastructure is paid for and the analytical value is unused.

03

One video analytics platform for multiple retail operations

Metrici analyses video from compatible IP cameras and turns it into structured events, counters and records. The same platform covers customer flow inside the store and vehicle movements outside it, so operations, marketing, facilities and security work from one data source instead of five separate tools.

Depending on the application and configuration, Metrici can:

  • Count people entering, leaving and passing a defined line or area
  • Track movement and activity levels across defined zones over time
  • Monitor occupancy and queue length against thresholds you configure, and raise alerts
  • Detect whether each parking space is free or occupied, by space and by level
  • Recognise vehicle number plates and apply whitelist or blacklist rules
  • Control connected barriers, LED displays, traffic lights and I/O controllers
  • Produce operational reports and expose events to external systems through documented interfaces
  • Run on your own servers, your virtualisation, or Metrici supplied hardware
Shopping centre atrium seen from above, with shoppers rendered as motion trails along the walkway between storefronts — the kind of overhead view used to measure customer flow.
What the camera seesAn elevated walkway view like this is what a counting line or zone is defined on: direction of travel, activity per area and total movement through the space.
Why it consolidates

Applications share one engine and one interface

Retail applications are modules on the same analytics engines, managed through the Metrici Interface. Adding queue monitoring to a site that already counts visitors is a configuration and licensing step, not a second vendor and a second console.

How AI video analytics compares to the alternatives

Each approach answers a different question. Video analytics is not a replacement for a POS system or a Wi-Fi sensor; it measures what happens in the physical space, which those systems cannot see.

What you need to know AI video analytics (Metrici) Manual counting Wi-Fi / Bluetooth sensing POS transaction data
People entering, continuous and unattended Yes — 24/7, per entrance or line Sampled, staff-dependent Estimated, opt-in dependent Only people who bought
Movement and activity by zone inside the store Yes — zones defined on the camera view Not practical Coarse, requires many sensors No
Queue length and wait-time pressure per lane Yes — with configured thresholds and alerts Spot checks only No Retrospective only
Free or occupied state of each parking space Yes — per space, per level Patrol-based, not continuous No No
Vehicle identity for access control Yes — plate recognition and rule lists Guard checks, manual No Payment ticket only, not identity
Uses cameras you already operate Yes, where cameras meet application requirements Not applicable Requires new sensor hardware Not applicable
Measures sales conversion No — combine with your own POS data No No Yes — within its own scope

The last row is deliberate. Metrici reports how many people entered a space and how they used it. Turning that into a conversion rate means joining it with your own transaction data — a reporting task on your side, not a metric the cameras produce.

04

Retail applications, grouped by the operation they serve

Deployments usually start with one application and add others on the same cameras and servers, where technically suitable. The six groups below describe what Metrici is configured to measure and what it can then do with that information.

Shopping malls and retail parks

Entrance and zone footfall, multi-level car park guidance, tenant and delivery access, reserved-bay monitoring across large sites with thousands of spaces.

Supermarkets and stores

Entrance counting, queue length per lane, occupancy of defined areas, self-checkout support and trolley or asset presence in the sales area.

Retail chains and commercial facilities

Consistent counters and reports across locations, employee access by plate, loading-area and yard control, supplier and contractor vehicle management.

A

Customer counting and flow analysis

How many people come, through which entrance, and how do they move through the space?

  • People entering and leaving across a counting line at an entrance, corridor or store front
  • Activity levels inside defined zones, tracked continuously rather than sampled
  • Time-of-day, day-of-week and monthly histories you can compare between entrances and sites
  • Occupancy of an area at any moment, derived from in and out counts
Uses Area Counter Line Counter MultiController
B

Customer flow, queues and occupancy

Where is service degrading right now, and by how much?

  • Queue length and people count per checkout lane, self-checkout bank or service desk
  • Configured thresholds that raise alarms to operators or to connected displays
  • Occupancy levels per area, useful for comfort, safety planning and capacity limits
  • A measured record of peak periods, so rota and lane-opening decisions can be compared to outcomes
Uses Area Counter Display output Custom rules
C

Parking space occupancy management

Which space is free, on which level, right now?

  • Free or occupied state for each individual space, not just a gate total
  • Availability per level, zone and direction, feeding LED totems and guidance displays
  • Session times per space, so dwell time and turnover become measurable
  • Reserved, disabled or tenant-only bays monitored for occupation, with plate verification where LPR is deployed
Uses Parking Place Detector Display output LAN Controller
D

Vehicle access and LPR control

Who is allowed in, and how do they get through without stopping?

  • Number plate recognition at entry and exit, with whitelist and blacklist rule lists
  • Automatic barrier, boom gate or traffic-light release for recognised vehicles
  • Tenant, employee and supplier groups with their own permissions and allocated spaces
  • A time-stamped record of every entry and exit, including denied attempts
Uses Metrici LPR LPR Parking LAN Controller
E

Delivery, logistics and yard areas

Are docks and yard areas moving, and who used them?

  • Plate-identified entry and exit for supplier and service vehicles at controlled gates
  • Access restricted to whitelisted vehicles and configured time windows
  • Vehicle presence and movement in loading areas and yard zones, as recorded by site cameras
  • Records usable for dock scheduling review and for supplier performance discussions
Uses Metrici LPR MultiController Transport & logistics
F

Security, alarm and situational awareness

What should someone look at now, rather than after the fact?

  • Configured alarms raised when a monitored condition is met in a defined area
  • Occupancy or crowd-density changes detected without an operator watching a wall of screens
  • Vehicle and plate events available to security and control-room staff in the same interface
  • Exportable event history that supports incident review and internal reporting
Uses Area Counter Parking Place Detector Smart City applications

What these applications do not do. Metrici does not identify individual shoppers and does not build personal profiles. It does not measure conversion rate — that requires joining visitor counts to your own transaction data. It does not issue fines on its own: parking and access applications support monitoring and documentation, while any sanction remains a human and procedural decision under the rules of the site and the applicable law. Predictive features, where configured, extrapolate from recorded history; they are not a forecast of demand.

Interior of a shopping centre with escalators, walkways and shop fronts, the kind of multi-zone environment where retail video analytics is deployed.
Mall and retail parkMulti-level layouts with several entrances and many distinct zones are where zone-level activity data changes decisions most.
Aerial view of a multi-level retail car park deck with marked parking spaces and access lanes.
Retail car parkOpen decks and multi-level car parks are typical Parking Place Detector environments: per-space state feeding availability guidance.
Fuel forecourts and convenience-at-petrol sites

Fuel retail has its own set of requirements — forecourt lane events, pump-area activity and canopy-environment cameras. Those are covered on a dedicated page.

Gas station applications
Parking monitoring and evidence capture

Where a site needs space-by-space records and vehicle images to support its own parking administration process, see the dedicated parking application page for what is captured and how it is handled.

Parking application
05

From existing camera feed to actionable operation

Four stages, in the order a retail deployment actually happens. Each stage is scoped per site, because camera suitability and the actions you want depend on the environment.

STEP 1

Capture

Compatible IP cameras already installed on site — or added where coverage gaps exist — provide the video feed. Resolution, framing, mounting height and lighting are assessed per camera against the requirements of each application.

STEP 2

Analyse

Metrici analytics engines process the streams and produce structured events: people crossing a line, occupancy in a zone, a space becoming occupied, a recognised plate. Processing runs on your own infrastructure, inside your network.

STEP 3

Monitor and manage

The Metrici Interface presents counters, occupancy, alarms, plate records and reports. Users see only the sites, cameras and records their role and tenant scope allows. Thresholds and rules are configured here.

STEP 4

Integrate and act

Events drive connected barriers, LED displays and controllers directly, and are published to access control, payment, reporting or BI systems through documented interfaces where an interface exists.

Camera suitability is confirmed before licensing, camera by camera. An existing camera that cannot deliver the image quality an application needs will be identified at this stage rather than discovered after go-live.

Metrici Interface people counter report with daily bar chart, and sidebar navigation for occupancy per day, occupancy per month, counters, alarms and object flow.
Stage 3 · The Metrici InterfaceDaily people counters with occupancy statistics and alarm history in the same interface operators use day to day.
Stage 2 · Where it runs

Your servers, your network

Analytics engines run on customer-supplied servers, on customer virtualisation, or on Metrici supplied hardware such as the Metrici Racksys series. Video does not have to leave your network to be analysed, which simplifies both security review and connectivity at large sites.

Stage 4 · Integration

Documented interfaces, not black boxes

Events, counters and plate records are exposed through documented interfaces so they can be consumed by access control, parking payment, ticketing, ERP or reporting platforms. Integration work is scoped per project against the interfaces that actually exist on the third-party side.

Rollout pattern

One application, one site, then extend

Most retail projects start with a single measurable application on a limited camera set, verify the results against what staff observe, then add further applications or locations on the same platform.

06

What retail operators gain operationally

The value is not the analytics itself. It is having a continuous, comparable record of how your retail space and its car park are actually used, so decisions can be defended afterwards.

✓

Continuous measurement instead of estimates

Visitor counts, occupancy and space state are recorded around the clock, in the same way every day. Seasonal patterns and the effect of a layout or promotion change become visible as a data series rather than an opinion.

✓

Better staffing and service decisions

Queue length per lane and measured peak periods give supervisors a basis for opening lanes and scheduling shifts, and a record of whether the change improved the situation.

✓

Better parking guidance and visibility

Per-space and per-level availability feeds guidance displays that drivers believe. Less internal congestion, shorter search time, and a better first and last impression of the site.

✓

More consistent vehicle access procedures

Tenant, employee and supplier vehicles recognised by plate pass on configured rules. Guard time is freed, entry records are consistent, and exceptions are logged with a plate and a timestamp.

✓

More value from cameras you already own

The same cameras that serve security also produce operational data. Retail analytics becomes an additional use of existing infrastructure rather than a parallel system with its own sensors and console.

✓

One platform across several operations

Footfall, queues, parking and access are configured and reported together. Operations, marketing, facilities and security look at one set of records instead of reconciling four tools.

✓

Consistent reporting across locations

Retail chains and multi-site operators apply the same definitions of a counted entrance or a monitored zone across sites, so locations are comparable rather than each measuring differently.

✓

Data stays under your control

Processing and event storage can remain on infrastructure you control, with permissions scoped by role, site and tenant. The retail operator sets retention and access; Metrici provides the controls to enforce those decisions.

A note on results. This page does not quote percentage improvements in sales, conversion or dwell time. Outcomes depend on your site, camera coverage and, above all, what your teams do with the data. The verified deployments below show what was installed and measured; all case studies are listed separately.

07

Retail applications and deployment examples

Three verified Metrici deployments, described by what was installed and what the system now records. They are not identical to every mall or store — a retail park with 26 pavilions and a business campus solve different problems — but they show the platform working at retail scale.

Street-level view of the Expo Market Doraly retail and wholesale park, showing pavilion buildings, a tenant directory totem and the main vehicle access area. Retail / wholesale park

Expo Market Doraly

Afumati, Ilfov · retail and wholesale park

Site22 hectares, 57,000 m² of retail space, 26 pavilions, more than 850 companies
Challenge3,000 parking places across a very large site, with vehicle access and visitor statistics managed manually
Installed8 Hikvision 3 MP LPR cameras at access points and one Metrici Racksys 240 server
Now recordsPlate-identified vehicle access, daily traffic analytics, custom visitor reports by county and by visit frequency
3,000PARKING PLACES 26PAVILIONS 850+COMPANIES 8LPR CAMERAS
Read the Doraly case
The Hermes Business Campus office building in Bucharest, with the lamp posts and access control bollards of the site car park in the foreground. Commercial campus

Hermes Business Campus

Bucharest · office and commercial campus (not a retail site, shown for tenant access management)

SiteMore than 2,000 parking spaces over roughly 78,000 m² across four access points, BREEAM Excellent
ChallengeSpace allocation and access for 30+ tenants without a central whitelist bottleneck
InstalledLPR-based access with hierarchical tenant groups, allocated space quotas and self-managed whitelists; integrated by ROLF
Now recordsEvery entry and exit per tenant, direction, time inside the site, and tenant activity reports per group
2,000+SPACES 30+TENANTS 4ACCESS POINTS
Read the Hermes case
Overhead site camera view of the Citygo car park in Bucharest, showing parked cars, an access lane, an outdoor seating area and the service building used for the car wash and tyre service. Commercial car park

Citygo, Bucharest

Bucharest · paid car park with additional vehicle services

SiteCommercial car park offering three services: parking, car wash and tyre service
ChallengeBilling three different services at different rates, and measuring dwell time without manual timing
Installed4 cameras with LPR covering entry and exit, multi-rate billing rules and automated dwell-time calculation
Now recordsPlate-identified arrival and departure per session, with duration feeding billing across all three services
3SERVICES 4CAMERAS MultiRATE BILLING
Read the Citygo case
See the systems in motion before you commit to a scope

Rather than a generic promotional clip, we prefer to walk your own environment: which camera views answer which question, and what the interface shows for them. Ask for recorded footage of people counting, occupancy alerts, space detection and plate-based access, or for a session against your camera specifications.

Request footage & a demo
Large-site vehicle access and tenant parking at scale

If your environment is closer to a logistics or truck park than a shopping centre — gated lorry spaces, driver check-in, long-stay monitoring — that deployment pattern is documented separately.

Truck parking applications
08

Products and integrations behind each application

A retail deployment combines one or more analytics modules with the control hardware and interfaces the site needs. Licensing is per application and per camera, scoped during the project.

Metrici component What it does in a retail deployment Typical applications
Area Counter Counts people in defined areas and zones, tracks activity and occupancy over time, and raises alarms on configured conditions. Footfall and flow analysis, zone activity, occupancy and queue monitoring, situational awareness
Line Counter Counts people or vehicles crossing a defined line, with direction, producing entrance and corridor totals. Entrance counting, store front traffic, corridor and zone boundary totals
Parking Place Detector Determines whether each individual parking space is free or occupied, per level and zone, and tracks session duration per space. Car park guidance, availability displays, reserved-bay monitoring, turnover analysis
Metrici LPR Recognises vehicle number plates and applies whitelist or blacklist rules, producing a time-stamped record of every read. Vehicle access control, tenant and employee parking, delivery gate control, session billing
LPR Parking Combines plate recognition with parking session logic for controlled entry and exit and duration-based records. Paid retail car parks, tenant parking, multi-service sites
Metrici MultiController Interfaces multiple cameras and devices into one control point, coordinating counters and outputs across an area. Multi-entrance counting, multi-lane checkout areas, gates with several cameras
Metrici LAN Controller Drives barriers, boom gates, traffic lights and displays over LAN and I/O connections when an event condition is met. Barrier release on plate match, LED guidance output, gate and dock control
Metrici Display Presents live availability, counter or status information on LED totems and screens for drivers, staff or visitors. Car park availability guidance, queue or status information at entrances
Metrici Custom Applies custom rules and reports when the required logic is not covered by a standard module. Site-specific visitor reports, unusual zone definitions, bespoke alarm logic
Custom applications Engineering work for requirements outside the standard catalogue, including bespoke integrations and reporting. Chain-wide reporting definitions, proprietary business system integration, unusual camera geometry
Cameras

Existing IP cameras

Works with compatible IP cameras meeting the resolution, framing and lighting requirements of the chosen application. Camera suitability is verified per camera before licensing, and cameras can be added where coverage is missing.

Outputs

Barriers, displays and controllers

Connected boom gates, traffic lights, LED totems and I/O devices are driven directly by event rules, so recognised vehicles and availability figures act without an operator in the loop.

Systems

Access control, payment and BI

Events, counters and records are published through documented interfaces so third-party access control, parking payment, ERP or reporting platforms can consume them where an interface exists.

09

Deployment and data governance

Retail environments are public spaces, so how a system is deployed and governed matters as much as what it measures. These are the questions procurement and legal teams actually ask, answered plainly.

Can Metrici use our existing retail cameras?

In many cases yes. Metrici works with compatible IP cameras that meet the image quality, resolution, framing and positioning requirements of the selected application. Suitability is assessed camera by camera during deployment planning, before licensing, so a camera that cannot support an application is identified early.

Can Metrici run on our own infrastructure?

Yes. Analytics engines run on server hardware on the customer's premises, on customer virtualisation, or on Metrici supplied servers. Video processing and the event database can stay inside the retail operator's own network, with no requirement to send video to an external cloud service.

Can Metrici connect to our existing access control or parking systems?

Metrici exposes events and records through documented interfaces and can drive connected barriers, displays and controllers directly. Integration with third-party access control, payment and business systems is possible where a documented interface exists, and is scoped as integration work during the project.

How are users, permissions and records managed?

The Metrici Interface supports named users, role-based permissions and hierarchical tenant or site scoping, so a tenant or store manager sees only the locations, cameras and records assigned to them. Plate events, corrections and exports are recorded, producing an audit trail of who accessed or changed what.

What video and event data is retained?

Retention is configured by the retail operator. Analytics events, counter totals and occupancy records are stored as structured data; video and vehicle images follow the retention period set in the recording systems. What is retained, for how long and who may access it are project decisions, documented before go-live.

Principle 1

Measure spaces, not people

Retail applications are configured to count and monitor activity in defined areas. They are not configured to identify individual shoppers, follow a person between visits or build personal profiles. Configuration choices that would move a deployment towards personal data must be assessed by the operator before they are enabled.

Principle 2

Least privilege by role and tenant

Access to live views, records and exports is granted per role and per site or tenant group. In multi-tenant retail and campus environments this lets each tenant administer its own vehicle list without seeing other tenants' data.

Principle 3

Human decisions stay human

Alerts and records support an operator's decision. Where a site uses parking or access records in an administrative process, the decision and any sanction remain with the responsible person under the site's rules and applicable law — the system documents, it does not adjudicate.

Metrici LPR Interface list of recognised plate events with columns for tenant, location, direction and time inside the site, illustrating scoped records and audit history.
Audit trail in practicePlate events with tenant, location, direction and duration — the record structure that makes access administration reviewable rather than dependent on guard notes.

Responsibility for compliance stays with the operator. Deploying Metrici does not by itself make a site compliant with GDPR or any other regulation. Lawful basis, notices and signage, data protection impact assessment, access policies, retention periods and cross-border transfers remain the responsibility of the data controller. Metrici provides technical capabilities — scoped permissions, structured event data, on-premises processing, configurable retention — that support a compliant deployment, and our approach to personal data handling is documented separately. Contractual terms are agreed per project.

10

Make your retail operations more intelligent

Start with the operation that costs you the most uncertainty today — footfall you cannot prove, queues you learn about too late, a car park nobody can guide, or a gate that needs a guard. We will map it to camera views and show you what the interface returns.

Request a retail demo Talk to our team

Or write to us directly: sales@metrici.ro

Bring three things

Site layout or floor plan, your camera inventory with models and positions, and the two or three operational questions you want answered. That is enough to scope which applications fit.

What you get back

A view of which camera positions support which measurements, the applications that apply, and what the Metrici Interface would show your teams for each one.

All case studies All applications How the AI works

Metrici AI Solutions SRL is a Bucharest-based company, founded in 2010. Metrici is our brand name for advanced video analytics solutions, designed to meet the needs of businesses and organizations. Our passion for technology and expertise in computer vision are used by the best partners for making a better place to live in, a safer and smarter city. With one of the widest range of applications based on video analytics and AI technologies, Metrici ranks among the leading companies in the world. Be assured that every software release represents our most advanced development to date as Metrici keeps innovating in accurate recognition.

recent post

Metrici goes to Romanian Security Fair

Metrici goes to RSF

October 5, 2026
Metrici Multicontroller version 3 with same Wiegand, RFID capabilities and 2 digital in, 2 digital out

Multicontroller version 3

July 21, 2026
Metrici lauches smart parking terminals for easy parking QR code access and cashless payment

Smart Parking Terminals Launch

June 26, 2026
Metrici LLM Server

Launch: Metrici LLM Server

May 18, 2026
Metrici Certind LLM

Certind automate certificate management

April 14, 2026
Metrici version 4.0, Ubuntu Support

Metrici 4.0, Ubuntu

February 10, 2026

Quick links

address

Vasile Lascar Street, 120, district 2, 020506, Bucharest, Romania

phone

+40742880099

Email

sales@metrici.ro

website

www.metrici.ro