Dynamic Web Lab

AI-Powered Video Analytics

AI Video Intelligence for Dubai Smart Cities & UAE Retail

Real-time CCTV analysis for Dubai and the UAE. YOLOv8-powered theft detection, zone monitoring, and behavior analysis. Open source, edge-ready.

View on GitHubRequest Demo
Object Classes
80+
Real-Time Processing
30 FPS
Edge-Ready
ONNX
Open Source
MIT

Powerful Detection Capabilities

Six core modules powered by YOLOv8 and ByteTrack for comprehensive video intelligence.

Person Detection & Counting

Accurately detect and count people in real-time using YOLOv8. Monitor foot traffic, occupancy, and crowd density across zones.

Theft Detection

AI-driven theft detection identifies suspicious behaviors and missing items. Instant alerts to security personnel.

Zone Monitoring

Define custom zones and get real-time counts of people entering, exiting, or loitering in restricted areas.

Behavior Analysis

Pose estimation and action recognition to detect unusual behaviors like fighting, falling, or running.

Face Detection

Detect and recognize faces in video streams. Enable access control and person identification workflows.

Edge Deployment

Export models to ONNX for running on edge devices. Low-latency inference without GPU dependency.

Built for Real-World Security

Deploy Rasd across industries to automate surveillance and improve response times.

Retail Security (Dubai Malls)

Prevent shoplifting, monitor store layouts, and optimize product placement in Dubai malls and UAE retail outlets.

Smart Cities (Dubai & Abu Dhabi)

Manage traffic flow, monitor public spaces, and enhance urban safety across Dubai and Abu Dhabi with AI-powered surveillance.

Public Safety (UAE Events)

Detect crowd disturbances, unattended bags, and emergency situations at UAE exhibitions, events, and public venues.

Building Security (UAE Offices)

Control access, detect intruders, and monitor restricted areas in Dubai office complexes, free zones, and facilities.

How It Works

From camera feed to actionable insights in four steps.

  1. 01

    Connect Camera

    Feed RTSP streams, MP4 files, or any video source into Rasd.

  2. 02

    AI Analyzes

    YOLOv8 processes each frame for real-time object detection and tracking.

  3. 03

    Detect & Alert

    Get instant alerts for theft, zone violations, unusual behavior, and more.

  4. 04

    Dashboard & Reports

    Review analytics, export reports, and monitor activity on a unified dashboard.

Processing Pipeline

A modular architecture from video ingestion to result delivery.

  1. 01
    Video InRTSP / MP4 / Webcam
  2. 02
    YOLOv8Object Detection
  3. 03
    ByteTrackMulti-Object Tracking
  4. 04
    Person Re-IDIdentity Matching
  5. 05
    Theft DetectionBehavior Analysis
  6. 06
    Zone CounterArea Occupancy
  7. 07
    ResultsAPI / Dashboard

Open Source — MIT License

Free & Open Source

Rasd is fully open source under the MIT license. Deploy it yourself at no cost, or get enterprise support for production environments.

Get Started on GitHubEnterprise Support

Frequently Asked Questions

Everything you need to know about getting started with Rasd.

What video formats does Rasd support?

Yes. Rasd supports RTSP streams, MP4 files, and any video format compatible with OpenCV. It can process live CCTV feeds via RTSP or uploaded video files.

Do I need a GPU to run Rasd?

No. Rasd runs on CPU with YOLOv8n (nano model). For higher accuracy, a GPU with CUDA support is recommended but not required. Rasd also supports edge deployment via ONNX.

Can Rasd integrate with existing CCTV systems in Dubai?

Yes. Rasd connects to any RTSP-compatible CCTV camera. It works with Hikvision, Dahua, Axis, and other brands commonly used in Dubai and UAE buildings.

Is Rasd production-ready for UAE businesses?

Yes. Rasd is MIT licensed, production-ready, and deployed in retail and smart city environments. It includes a web dashboard, API, and database storage.

Built With

  • Python 3.12+
  • YOLOv8
  • FastAPI
  • SQLite
  • PostgreSQL
  • ONNX