Skip to content
SWAT Alliance logo
Request a quote

Methodology guide

Traffic Video Analysis Methodology

A controlled process for converting suitable project video into traceable traffic observations.

Technical visual · Traffic video analysis workflow
Traffic video analysis workflowOBSERVECLASSQA

Definition and purpose

Traffic video analysis is human-led observation of approved footage against a written data specification. It can produce agreed volumes, movements, classes, queues or active-travel observations when the source material supports them.

Use it when distributed locations, repeatable review or access constraints make footage an appropriate observation source.

Required inputs

  • Securely shared project footage
  • Site plan and observation boundaries
  • Dates, time periods and time zone
  • Movement, class and interval rules
  • Delivery and data-handling requirements

Video and camera considerations

Check stability, clock alignment, visibility, glare, weather, occlusion, resolution and the full observation area before accepting footage. A camera does not compensate for an unobservable movement.

Observation, classification and intervals

Observation process

Review footage against the approved observation matrix, record only the requested events, and keep a project note for source gaps or approved exceptions.

Movement and classification process

Use definitions that can be applied reliably from the footage. Escalate ambiguous instances using the project rule rather than assuming body type or axle count.

Time intervals

Source timestamps and delivery intervals should be confirmed before processing; continuity checks identify gaps and duplicated periods.

Validation and QA/QC

Review coverage, timing, totals, movement logic and class consistency. Retain documented exceptions with the delivery package where agreed.

Common problems and limitations

  • Glare, darkness and compression can hide vehicle features.
  • Clock drift can misalign multiple cameras.
  • Overlapping road users may conceal an individual movement.

Video analysis cannot recover detail that is not visible in the footage and should not be described as fully automated unless that is demonstrably true for the supplied workflow.

Typical deliverables

  • Quality-reviewed interval data
  • Movement/classification tables
  • Method and source notes
  • Exceptions or coverage notes

Illustrative output format

Illustrative sample data only — not a project result or benchmark.

CameraWindowObservationStatus
Illustrative camera A07:00–09:00Movement matrixReviewed

Frequently asked questions

Can large video files be uploaded through the public form?

No. The form is for a project brief; approved secure-transfer arrangements are confirmed separately.

Can one video support several outputs?

Potentially, if all requested areas and rules are visible and agreed before processing.