Edge-to-Cloud Architecture

Proactive Surveillance
Powered by AI

Beyond traditional CCTV. CloudVision AI doesn't just record; it understands. From detecting structural fires to identifying wild animal intrusions and unauthorized human access, it keeps you safe in real-time.

Comprehensive AI Protection

Trained to recognize multiple threats with high confidence.

Fire & Smoke Detection

Our custom-trained edge AI identifies flames and smoke the moment they appear, minimizing false alarms from reflections and triggering instant emergency protocols before a fire spreads.

Wildlife Intrusion

Protect farms, rural estates, and remote facilities. CloudVision recognizes wild animals (e.g. bears, boars, coyotes), alerting you before they pose a threat to livestock, property, or human life.

Unauthorized Access

Go beyond dumb motion sensors. Smart human-detection filters out shadows, pets, and wind to alert you only when a genuine security breach or perimeter intrusion occurs.

Edge-First Processing

Video analysis happens locally on the camera edge (using YOLO). This ensures zero latency and absolute privacy, while only sending critical event data to the cloud.

Multi-Channel Alerts

When seconds matter, we don't rely on just one channel. Get notified simultaneously via WhatsApp, Email, and a dedicated, high-priority Android Pager app that sounds an alarm.

Smart Pre/Post Buffering

Never miss the context of an event. While monitoring, a rolling buffer caches footage. Upon detection, it saves the crucial 10 seconds before and after, providing a complete picture.

How It Works

From camera lens to your phone in under a second.

1. Continuous Edge Monitoring

The Edge Engine runs a highly optimized AI model locally. It analyzes every frame of the live camera feed in real-time, functioning completely independent of internet stability for its core analysis.

2. Intelligent Event Trigger

Upon detecting an anomaly (like a fire breaking out, or an animal entering the frame), the system validates the threat. It instantly saves a pre/post-event video clip from its internal rolling buffer.

3. Secure Cloud Sync

The critical video clip and a high-resolution snapshot are securely transmitted to the CloudVision FastAPI backend, where they are logged, stored, and prepared for rapid broadcast.

4. Instant Fan-Out Alerting

The backend triggers simultaneous alerts across all platforms. The web dashboard updates via WebSockets, stakeholders receive WhatsApp/Email messages, and the Android Pager rings loudly until acknowledged.

Meet Our Team

The innovators behind CloudVision AI.

Pratik Lohar

Pratik Lohar

Developer

LinkedIn
Anushka Deshmukh

Anushka Deshmukh

Research & Cloud Operations

LinkedIn
Manasi Thorat

Manasi Thorat

Research & Cloud Operations

LinkedIn
Anushka Shirke

Anushka Shirke

UI/UX

LinkedIn

Acknowledgments

With immense gratitude to our mentors.

We would like to express our deepest gratitude and special thanks to our Guide Prof. G. P. Mohole Mam and our Co-ordinator Prof. S. P. Zope Mam. Their continuous support, invaluable guidance, and encouragement were instrumental in bringing this project to life.