Skip to content

eliaghazal/crashlens

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 

Repository files navigation

CrashLens

AI-powered crash detection and emergency response platform

Website Status Stage


What is CrashLens?

CrashLens is a real-time crash detection and emergency coordination platform that bridges the moment of impact to the people who need to respond — in seconds, not minutes.

A Raspberry Pi 5 edge device reads live IMU, GPS, camera, and LTE data. When a crash is detected, the system instantly:

  1. Captures a 15-second pre-crash + 30-second post-crash video clip
  2. Uploads GPS coordinates, speed, and impact severity to the cloud
  3. Notifies the insurance dashboard with AI forensic analysis of the incident

"From sensor to responder in under 3 seconds."


Proven End-to-End

CrashLens has been tested and verified on real hardware:

Component Status
Camera Module 3 NoIR (1080p30) Proven
GPS Module (NEO-6M, UART) Proven
IMU Crash Detection (100Hz, 2.5G threshold) Proven
Ring Buffer Recording (15s pre + 30s post) Proven
H264 → MP4 Conversion (ffmpeg) Proven
Cloud Upload (metadata + video to R2) Proven
Heartbeat Monitoring Proven
Systemd Auto-Start on Boot Proven
Insurance Dashboard (8 pages) Proven

Architecture Overview

+------------------------------------------------------------------+
|                     EDGE DEVICE                                   |
|  Raspberry Pi 5 · IMU/GPS/Camera/LTE                             |
|  Real-time crash detection · AI inference on-device              |
+----------------------------+-------------------------------------+
                             |
                          LTE/HTTPS
                             |
                             v
+------------------------------------------------------------------+
|                    CLOUD BACKEND                                  |
|  Cloudflare Workers + Hono.js (edge API)                         |
|  Supabase (Postgres + Auth)                                       |
|  Cloudflare R2 (video evidence storage)                           |
|  Multi-stakeholder routing                                        |
+--------------------+--------------------------+------------------+
                     |                          |
             +-------+------+          +--------+-----------+
             | Flutter App  |          | Next.js Dashboard  |
             | Driver +     |          | Insurance / Fleet  |
             | Responder    |          | Management         |
             +--------------+          +--------------------+

Tech Stack

Layer Technology
Edge Device Raspberry Pi 5, Python, OpenCV, TensorFlow Lite
Connectivity LTE (SIM7600G), GPS (NEO-6M), 6-axis IMU
Cloud API Cloudflare Workers + Hono.js (TypeScript)
Database Supabase (PostgreSQL + Row Level Security)
Video Storage Cloudflare R2
Mobile Flutter (iOS + Android)
Dashboard Next.js 14, TypeScript, Tailwind CSS
AI On-device crash inference, cloud forensics pipeline

Status

CrashLens is a live, deployed commercial product. The platform is currently in active development and pilot testing.

Visit: crashlens.org


Team

Elia Ghazal — Co-Founder & Lead Engineer GitHub · [email protected]

Collaborators


Commercial product — source code is proprietary. This repository is a public showcase only.

About

CrashLens — embedded vehicle crash detection + AI-assisted forensic evidence platform. IEEE ICCA 2026 (under re-review). Public showcase.

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors