Inspiration
During house or apartment fires, emergency responders prioritize saving human lives. Unfortunately, pets trapped inside are often overlooked simply because rescuers have no way of knowing they are present. We were inspired to create a solution using computer vision that ensures no pet is left behind during an emergency.
What it does
Animal Pet Rescue is an AI-powered monitoring system that continuously analyzes live video streams using two dedicated YOLO models:
- Animal Detector: Identifies pets inside the residence.
- Fire & Smoke Detector: Detects visual signs of fire or smoke.
When both conditions occur simultaneously (Fire + Animal), the system triggers a smart emergency alert and instantly sends a notification via Telegram to the pet owner and local emergency responders, complete with animal details and a live photo screenshot.
How I built it
The system was developed using Python as the core programming language. OpenCV was used for real-time video stream processing and frame extraction. We deployed YOLO (You Only Look Once) models to perform dual-detection for animals and fire/smoke. Finally, we integrated the Telegram Bot API to deliver real-time photo alerts and data instantly to users.
Challenges I ran into
- Optimizing dual YOLO model execution simultaneously on a live stream without causing significant frame lag or delay.
- Ensuring logic accuracy to eliminate false alarms so that alerts trigger strictly when fire/smoke AND an animal are present in the frame together.
Accomplishments that I'm proud of
- Successfully building a working end-to-end alert pipeline that captures a live frame and delivers it to Telegram within seconds.
- Creating an accessible, low-cost safety technology that has the potential to save lives.
What's next for Animal Pet Rescue
- Integrating thermal imaging to detect pets hiding in dense smoke.
- Adding Direct SMS/Call notification capabilities for areas with weak internet