Architecture

How Axix Surveillance works

Connect cameras, process every frame with CPU-only AI, deliver alerts instantly, and audit everything.

Connect your IP cameras — Axix Visual Intelligence
Step 1

Connect your IP cameras

Add any IP camera from the web dashboard — test the connection, configure per-camera AI settings, assign zones, and go live.

  • Dahua, Hikvision, and any IP/HTTP stream
  • Per-camera AI configuration
  • Zone assignment before go-live
AI processes every frame — Axix Visual Intelligence
Step 2

AI processes every frame

Person detection → video tracking → face recognition → zone check → alert generation. All on your CPU.

  • CPU-optimized models — no GPU
  • Multi-stage pipeline per camera
  • Real-time processing at 15–30 FPS
Alerts reach your team instantly — Axix Visual Intelligence
Step 3

Alerts reach your team instantly

Zone violations, blocked persons, intrusions, and camera failures delivered to web dashboard and mobile guard app.

  • Snapshot image on every alert
  • Mobile guard acknowledge workflow
  • Tiered INFO, WARN, CRITICAL feed
Review, audit, and report — Axix Visual Intelligence
Step 4

Review, audit, and report

Every detection, check-in, violation, and alert stored with images and timestamps. Search and export for compliance.

  • Search by person, camera, date, alert type
  • Cross-camera movement timelines
  • CSV and PDF exports
From detection to guard response — Axix Visual Intelligence
Event flows

From detection to guard response

Typical pipelines for blocked persons, zone violations, cross-camera tracking, and footfall counting — all logged and auditable.

  • Blocked person → alarm → mobile alert
  • Zone violation → snapshot → dispatch
  • Full event lifecycle audit trail
Deployment flow

Four steps to go live

01

Connect Your IP Cameras

Add any IP camera (IP camera, HTTP, Dahua, Hikvision, or others) from the web dashboard. Test the connection, configure per-camera AI settings, assign it to a zone, and go live. The system starts processing frames immediately.

02

AI Processes Every Frame

Each camera feed runs through a multi-stage pipeline: Person Detection (person detection) → Tracking (video tracking) → Face Recognition (face matching) → Zone Check → Alert Generation. All processing runs on your CPU — CPU-optimized models.

03

Alerts Reach Your Team Instantly

Zone violations, blocked persons, intrusions, and camera failures — all delivered in real time to the web dashboard and mobile guard app. Every alert includes a snapshot image and timestamp.

04

Review, Audit & Report

Every detection, check-in, violation, and alert is stored with images and timestamps. Search by person, camera, date, or alert type. Export CSV/PDF. View cross-camera movement timelines.

Event pipelines

Typical event flows

From camera detection to mobile alert — every path is logged and auditable.

Blocked Person Detected

Camera IP camera→AI Pipeline→Person detection→Face Recognition→face matching Match → BLOCKED→Alert Saved→real-time Alarm→Web Dashboard + Mobile App

Zone Violation

live video Worker→Person in Zone→Identity + Card Color Check→Whitelist FAIL→Violation Image Captured→Alert broadcast→Zone Gallery Updated→Mobile Alert

Cross-Camera Tracking

Camera A: Person detection → Tracking → appearance-based tracking→unified person ID Assigned→Camera B: Cross-camera appearance match→Same unified person ID→Timeline UI Updated

Daily Footfall Count

Entry Camera→Person Detected→Match against today's unique visitors→New Unique Visitor→Count Incremented→Dashboard Live Update
AI pipeline

Multi-stage processing on your CPU

IP Camera → Person Detection → Video Tracking → Face Recognition → Zone Check → Alert Generation

FAQ

Technical questions

Can guards use their own smartphones?
Yes. mobile mobile app runs on Android and iOS. Guards configure server IP, log in with admin-created credentials. Alerts scoped to their assigned area and types.
Zone violations vs no-go intrusions — what's the difference?
No-go areas = absolute exclusion, anyone triggers alarm. Zones = access-controlled with whitelist — only unauthorized persons trigger violation. Both can run on same camera.
Can it track someone wearing a mask or facing away?
Yes. appearance analysis (body shape, clothing, movement) work when face isn't visible. System fuses appearance-based tracking (60%), clothing color (30%), and face (10%) for robust matching.

Ready to make your cameras intelligent?

Request a demo, download the datasheet, or talk to our team about on-premise deployment for your facility.