IBSAR
Initialising vision engine…
Vision AI · running live on the edge

Give every camera
the power to understand.

IBSAR turns ordinary CCTV into an intelligent sensor network — counting people, recognising faces, and tracking vehicles in real time, processed on-site with state-of-the-art computer vision.

0Detection accuracy
0Frames / second
0Edge monitoring
expand_more
Deployed for education & enterprise across Palestine
Birzeit University Shkokani Group On-prem & Hetzner cloud YOLO11 · ArcFace · Qdrant Real-time edge inference
Live detection engine

This is what IBSAR sees.

A real-time simulation of our tracking pipeline — objects are detected, boxed, classified and counted the instant they cross the line. No cloud round-trip required.

LIVE CAM_01 · Main Entrance
-- FPS
directions_walk
0People
directions_car
0Vehicles
check_circle
0Total today
bolt
On-device inference

YOLO11 runs on the client machine — footage never leaves the building.

timeline
Multi-object tracking

ByteTrack keeps a stable ID on every subject through occlusion.

tune
Zone & line rules

Draw counting lines and zones; IBSAR does the rest, per camera.

One platform, three lenses

Purpose-built vision products

Three modules are live in production today — with more on the roadmap. Every module shares one dashboard, analytics and access layer.

groups
Deployed

People Counting

YOLO11 · ByteTrack

Count entries and exits in real time with directional line-crossing — live occupancy, footfall trends and hourly heatmaps.

  • checkDirectional in/out counting
  • checkLive occupancy & dwell time
  • checkMJPEG live streaming
directions_car
Deployed

Vehicle Analytics

Parking · traffic

The tracking core tuned for vehicles — count traffic, manage parking-slot occupancy and monitor gates with vehicle-specific models.

  • checkVehicle counting & class
  • checkParking slot occupancy
  • checkGate & barrier events
psychology
Coming soon

Behavior Prediction

Anomaly · intent

Go beyond counting — flag loitering, crowding and falls, and surface unusual movement as anomalies before they escalate.

  • boltAnomaly & loitering alerts
  • boltCrowd & fall detection
  • boltPredictive risk scoring
category
Coming soon

General Object Detection

Custom classes

Train IBSAR on the objects that matter to you — detect, count and track any custom class straight from your own footage.

  • boltCustom-trained models
  • boltAny object class
  • boltZone & count rules
confirmation_number
Coming soon

License Plate Recognition

ANPR

Automatic number-plate recognition to pair with Vehicle Analytics — log plates at gates and match against allow / deny lists.

  • boltPlate OCR at gates
  • boltAllow / deny lists
  • boltEntry / exit logs
The platform

Everything around the AI

A complete operations layer: dashboards, multi-tenant access, alerts and reporting.

insights

Real-time analytics

Live charts, trends and dashboards that update as events stream in.

videocam

Multi-camera

Manage every RTSP feed and site from a single pane of glass.

admin_panel_settings

Module-scoped RBAC

Fine-grained roles and permissions per client, module and page.

notifications_active

Smart alerts

Threshold and event-based notifications when it matters.

dns

Edge + cloud

AI on the client machine, aggregation on the IBSAR server.

history

Historical reports

Query and export any period with powerful filtering.

Architecture

From lens to insight in four steps

01
videocam

Camera

Any RTSP camera streams frames to the on-site engine.

arrow_forward
02
memory

Edge AI

YOLO11 detects, tracks and counts — locally, in real time.

arrow_forward
03
storage

Sync

Counts persist to PostgreSQL & InfluxDB, then sync to the server.

arrow_forward
04
dashboard

Dashboard

Insights, live feeds and reports in your browser.

Ready to let your cameras think?

Open the IBSAR dashboard and watch the vision engine work in real time.

rocket_launch Launch IBSAR