Internet of Tubes

ML traning data sample (No Leak)
ML training data sample (Leak)

Inspiration

To maintain large buildings and industrial machinery, it is important to identify chemical/liquid as early as possible. Our system provides a means to autonomously use robots to survey and notify if a leak is detected.

Workflow:

Control panel detects Gas leaks
A robot is dispatched to look for leak
Robot detects and reports leak

What it does

The system uses a combination of an IR camera, Google Cloud AutoML and MQ Gas sensors to detect pipe leaks.

How we built it

We build 3 components:

Handheld Thermal cam: Raspberry Pi 3 A+, Adafruit AMG8833, Pimori HyperPixel, Raspberry Pi Camera, Adafruit Boost 1000, Lipo Battery
Robot: Elegoo Robot Smart Car Kit
Control Panel Kiosk Raspberry Pi and MQ Gass sensors.