Inspiration

We were inspired by the high economic loss due to cracked surfaces in propriety materials and infrastructure. Many of these damages could have been avoided if caught earlier. So we invented CRACKD!

Description

CRACKD is a fully automated computer vision system to detect cracks and deformities in surfaces. Utilizing a convolutional neural network trained on TensorFlow coupled by real-time recordings with a Pi Camera, CRACKD can sort and classify metals and concrete surfaces as cracked or non-cracked within fractions of a second at high precision.

Kick Starter Demo

https://youtu.be/czKjyvaCqx0

Share this project:

Updates