Computer vision · Operations and logistics
See where the flowcomes to a halt.
Operational monitoring with computer vision measures, from camera images, how long each stage of the flow takes: how long a lorry stays at the dock, how long a pallet waits in the transit area, where machinery builds up. The data shows the bottlenecks, and the operations team decides what changes.
You know the flow gets stuck, but not where or when
Delays happen at the handover between stages: the lorry waiting for a dock, the pallet forgotten in the transit area, the forklift standing idle after unloading. Without measurements, the causes remain guesses, and improvements are made on intuition.
- Dwell times at the dock or in the yard are not measured systematically
- Pallets stay in transit areas longer than they should, without anyone noticing
- The WMS shows when something was scanned, not what physically happened between two scans
- Manual timing is rare, costly and influences the very behaviour being measured
Who it is for
- Logistics centres and warehouses with loading docks and transit areas
- Factories with material flows between production stages
- Operators of lorry yards, terminals or machinery depots
- Operations managers who want real times, not estimates
What the system can track in the warehouse and yard
- 01
Dock and zone occupancy
Shows at any moment which docks are occupied and how long each vehicle has been there.
- 02
Dwell times
Measures how long a vehicle, pallet or batch stays in a zone and flags when the set limit is exceeded.
- 03
Handover between stages
Records the moment a load enters and leaves each stage, to calculate durations.
- 04
Identifying bottlenecks
Highlights the zones and hourly intervals where most waiting builds up.
- 05
Correlation with systems
Places visual events alongside data from the WMS, TMS or ERP, to show the differences between the physical flow and the records.
How it works
From images to process decisions
The steps marked in yellow are approved by a person.
- Input
- AI
- Action
- Human in the loop
- Result
- 01
Input
Cameras by zone
Compatible cameras cover the docks, the yard and the transit areas.
- 02
AI
Detection and tracking
The model recognises vehicles, pallets and machinery and tracks them within the zone.
- 03
Action
Stage time calculation
Entries into and exits from zones become durations and waiting queues.
- 04
Action
Operations dashboard
Occupancy and times appear on a dashboard, with alerts when limits are exceeded.
- 05
Human in the loop
The team analyses causes
The operations manager interprets the data and decides on changes.
- 06
Result
Data-driven process adjustment
The effect of changes is tracked on the same measured times.
What data and video coverage are needed
- A flow map: the stages, the zones and what entering and leaving each one means
- Video footage from each zone, at normal and peak hours
- Labelled examples for the types of vehicles, machinery and load units tracked
- The time limits considered normal for each stage
- Access to data from the WMS, TMS or dock schedules, if correlation is wanted
- Sufficient video coverage of the zones, including at night, where shifts are worked
Possible integrations
- IP cameras with an RTSP stream
- ONVIF-compatible cameras
- NVRs and VMS systems
- WMS systems
- TMS and dock scheduling systems
- ERPs such as SAP or Microsoft Dynamics 365
- Power BI or Grafana dashboards
- MQTT
- REST API and webhooks
- IP cameras with an RTSP stream
- ONVIF-compatible cameras
- NVRs and VMS systems
- WMS systems
- TMS and dock scheduling systems
- ERPs such as SAP or Microsoft Dynamics 365
- Power BI or Grafana dashboards
- MQTT
- REST API and webhooks
Each integration's compatibility is checked during the analysis phase, for the systems and versions you use.
Control and transparency
What stays with people
- Interpreting the causes of a bottleneck and deciding to change the process, which stay with the operations team
- Setting time limits and reviewing them when volumes or the layout change
- Checking the data before it is used in discussions with hauliers or suppliers
- Making sure the system tracks vehicles and loads and does not assess individual employee performance
Limits and operating conditions
- Objects that leave the frame or are hidden by others can be lost, and that measurement is marked as incomplete
- Tracking across cameras works best when their views overlap; without overlap, doing it visually alone is difficult and sometimes impossible
- Rain, fog, snow and poor yard lighting can reduce detection quality
- Very similar pallets or vehicles cannot be told apart individually from the image without a visible marking
How we verify the measured times
We agree the indicators together before launch and track them afterwards. The values depend on the process, the data and the volume.
- Accuracy of times
- Calculated times compared with manual timings or gate records, on a sample
- Missed events
- Zone entries or exits the system did not record, found during video review
- Consistency with systems
- Differences between visual events and WMS or TMS data, analysed together with the team
- Usefulness for decisions
- How many of the flagged bottlenecks the team confirmed as real process problems
How we implement operational monitoring
- 01
We map the flow
We draw the stages and zones together with the operations team and agree which times matter.
- 02
We check the video coverage
We review existing cameras, stream compatibility and coverage gaps, by day and by night.
- 03
We configure the zones and the model
We adapt the model to your vehicles and loads and validate the times against manual measurements.
- 04
Dashboard and regular review
We launch the dashboard and regularly review, together with the team, what the data shows and what is worth changing.
Use cases
The scenarios below are conceptual examples, designed to illustrate realistic applications. They are not implementations delivered by WEB-AZ.
- Logistics
Dwell time at the docks
Cameras above the docks record the arrival and departure of each lorry, and the dashboard shows the docks where unloading consistently takes longer.
Conceptual example - Manufacturing
Pallets forgotten in the transit area
When a pallet stays in the area between production and the warehouse beyond the set limit, the shift supervisor receives a notification.
Conceptual example - Building materials
Yard congestion
An hour-by-hour analysis of the yard shows the intervals when lorries wait longest, and the team reviews the arrival schedule.
Conceptual example
Useful guides
Before you decide, read
Frequently asked questions
What is operational monitoring with computer vision?
It is measuring the physical flow from camera images: when a vehicle, pallet or batch enters and leaves each zone and how long it stays there. The result is dwell times, dock occupancy and bottlenecks by hourly interval. Unlike video alerts, the main purpose is process analysis, not reacting to an incident.
Do we monitor employees with this system?
That is not the purpose. The system tracks vehicles, pallets and machinery in zones. Any use that would assess individual employees must be justified, proportionate, communicated and legally reviewed beforehand.
Can we read number plates?
It may be technically possible, but number plates can constitute personal data. We treat this as a separate feature, with a clear purpose, information for those concerned and limited retention, and only after assessing whether it is necessary.
What happens when the system loses a vehicle from the frame?
That measurement is marked as incomplete, not filled in by guessing. The team sees how many events are incomplete and can review the footage, and repeated gaps show that a camera needs to be moved or added.
Where is the data stored?
Wherever possible, images are processed locally and only events and times are kept. If clips are also kept, we agree the retention period and access in line with your video surveillance policy.
Can it be linked to the WMS or TMS?
It is possible if your system offers an API or data exports. We check access to each system before confirming the integration.
Can’t find the answer? Write to us and we’ll reply directly.
Where do you think your flow gets stuck?
Describe the stages and zones where delays occur. We assess whether they can be measured visually and what video coverage would be needed.
Related solutions
- Video event detectionVisual rules that flag entries into restricted zones, missing protective equipment or fallen objects, for human review.
- People and object countingAnonymous counting of crossings, zone occupancy or the items in a batch, without identifying anyone.
- AI agent for operationsTracks internal requests, tickets and stock thresholds, prepares tasks for teams and coordinates handovers between departments, with a person responsible for every decision.