SentryAI – Smart Accident Sentinel

An AI-powered, real-time accident detection and reporting system using voice, image, and sensor data to prevent, track, and learn from incidents

Comment

Inspiration

We were inspired by the alarming frequency of preventable accidents in workplaces, roadways, and public environments. Traditional accident logging is often reactive, manual, and disconnected from real-time data. We wanted to create something proactive, intelligent, and human-centered that could not only track accidents but help prevent them before they happen

What it does

Our project, SentryAI, is an AI-powered accident tracker and prediction system. It combines voice logging, photo-based detection, and sensor triggers to:

Log incidents in real-time

Auto-classify type/severity using computer vision and NLP

Display patterns through interactive heatmaps and dashboards

Predict high-risk zones and times using historical data

It also generates instant safety reports and notifies the right teams when action is needed.

How we built it

We used the following stack:

Frontend: React with Mapbox for interactive heatmaps

Backend: Flask with MongoDB to store reports and sensor data

Integrations: Twilio for real-time alerts and Slack bot for team updates

We also simulated wearable sensor data to mimic real-world fall detection scenarios.

Challenges we ran into

Integrating multiple input modes (voice, image, sensor) into one unified backend

Fine-tuning the accident classifier to accurately identify incident types from photos

Creating a clean UI to visualize dense data while keeping it user-friendly

Time constraints—some predictive features were only partially trained

Accomplishments that we're proud of

Multimodal input (voice + image + sensor) makes safety reporting far more accessible and inclusive

AI can help reduce not just response time but accident rates with the right data

Clear, proactive communication systems can empower safer environments

What we learned

Multimodal input (voice + image + sensor) makes safety reporting far more accessible and inclusive

AI can help reduce not just response time but accident rates with the right data

Clear, proactive communication systems can empower safer environments

What's next for SentryAI – Smart Accident Sentinel

Add live wearable device integration (e.g., accelerometer fall detection)

Expand the predictive model with more real-world datasets

Open-source the project for NGOs, industries, and public safety use

Built With

Share this project:

Updates