Vehicle classification
Cars, trucks, buses and two-wheelers.
Technology
Traffic intelligence.
Using computer vision and mobile robotics to understand how roads are used, so people can manage congestion and plan better.
Concept
Capabilities
What traffic intelligence could measure.
Vehicle counting
Counts by lane, time and direction.
Traffic density
Occupancy and flow patterns across a road section.
Congestion analysis
Identifying where and when queues form.
High-resolution imaging
Detailed imagery for reliable detection.
Smart traffic reporting
Automated summaries for planners and operators.
How it could work
From camera to insight.
Capture
Cameras on a mobile platform
Detect
Find vehicles in each frame
Classify
Identify vehicle types
Aggregate
Count and measure density
Report
Dashboard and summaries
Number plate recognition
Demonstration research, handled carefully.
SAANMI is studying number plate recognition as a demonstration capability. Plate data can identify individuals indirectly, so any real deployment would require a lawful basis, clear purpose, data minimisation, retention limits and appropriate authority.
Privacy note. Camera-based systems may process personal data. See the draft privacy policy. Use by or for public authorities would be subject to applicable law and approvals.
Working on traffic or smart-city problems?
We welcome research and institutional collaborations.