Traffic Monitor: Vehicle Counting And Traffic Analytics Software
Many cameras, one server, one dataset. Counting and classification across an entire road network with historical time-series, not a per-intersection appliance for every approach.
Person and vehicle detection benchmarked on VisDrone, 93.4 mAP@0.5 · $48 per camera per month
What It Measures
Vehicle count by direction and by class
Flow rate over time
Congestion state
Historical time-series across the full camera network
LPR available as Addon
Traffic & Vehicle Analytics
Vehicle Detection & Classification
- Cars, motorcycles, buses, trucks
- Lane-based classification
- Directional movement tracking
Traffic Flow & Congestion Analysis
- Vehicle count per lane and direction
- Traffic density measurement
- Congestion detection and trend analysis
Reporting & Analytics
- Real-time dashboard access
- Export capabilities (PDF, Excel)
Why Choose the Software Edition
Ideal for smart city platforms, transportation authorities, and enterprises with established IT infrastructure. No servers at the roadside? The vehicle counting camera system runs the same analytics on an appliance.
Deploy on existing servers or private infrastructure
Maintain full data ownership and privacy
Integrate with existing ITS, dashboards, or data platforms
IMPLEMENTATION
How It Deploys
Central server
One deployment covering every camera that routes back to your control centre. Add approaches by adding licences, not by adding hardware.
Define counting geometry
Lines, directions, and class assignments per camera through the web interface.
Feed the platform you already run
REST, webhook, database export, or scheduled reporting into your traffic management system.
TRANSPARENT
Pricing
| Licence | Price |
|---|---|
| Subscription, per camera per month, billed annually | $48 |
| Perpetual, per camera, one time | $1,430 |
| Care on perpetual, per year, from year 2 | 18% of licence value, optional |
Four-camera minimum. Includes model updates and support.
MINIMUM
Technical Specifications
| Streams per GPU at traffic resolution | 2 to 3 streams per 6 GB GPU; scales linearly with memory footprint |
| Time-series storage sizing | Highly compressed metadata; ~100MB per camera per day |
| Export formats | CSV, PDF |
| API Integration | REST API, Webhook |
| Historical retention | Configurable by operator based on local storage allocation |
Integration
Detection results are only useful where your team already works. Every event is routable.
Push
- REST API payloads
- Webhook event push
- MQTT broker support
Pull
- CSV export
- Historical query API
- PDF reporting delivery
Identity & Access
- Role-based access control
- Immutable audit logging
- Granular camera permissions
FAQ
Software Questions
How many camera streams can one server handle?
A single 6 GB CUDA-capable GPU handles two to three streams. Twenty streams requires proportionally more GPU memory and compute; exact sizing depends on resolution and frame rate. Request a sizing sheet with your stream profile.
How long is historical data retained?
As long as you allocate storage for. Retention is configured by you; ARSA does not hold the data on self-hosted environments.
What export formats are available for reporting?
CSV, XLS, and PDF, plus API for direct integration capabilities.
Can it integrate with our existing ATMS?
Yes. Aggregated counts, class data, and event metadata are pushed via REST API or webhook directly into your ATMS or reporting stack.
Can one deployment cover multiple sites?
Yes. A single deployment covers every camera that routes back to it, across any number of intersections and corridors, under one dashboard.
How far can one camera count vehicles?
Counting needs 67 pixels per metre at the vehicle, up to 24.9 m on a 2 MP camera with a 60-degree lens. Classification needs 107 pixels per metre, up to 15.6 m, and licence plate reading needs 325, up to 5.1 m. A higher resolution sensor or a narrower lens extends each distance.
Which vehicle classes does it count?
Cars, motorcycles, buses and trucks, by lane and by direction, with flow rate, density and congestion over time across every camera in the deployment.
How much does vehicle counting software cost?
$48 per camera per month, billed annually, or $1,430 per camera as a perpetual licence with optional Care at 18% a year from year 2. There is a four-camera minimum, and model updates and support are included.
Do we need servers to run it?
Yes. It runs on servers you own, with one 6 GB GPU handling two to three streams. Sites with no server can run the same analytics on an AI Box at the roadside.
Deploy Traffic Intelligence on Your Infrastructure
Transform existing CCTV networks into data-driven traffic monitoring systems.