Hackathon ยท Computer Vision
HackFest 2023 : Crack Detective
A computer-vision pipeline that flags cracks and dents on large surfaces and measures their dimensions. Classical preprocessing pairs with machine learning inference to help inspectors prioritise fixes faster.
- OpenCV preprocessing normalises lighting and isolates contours before detection.
- Custom ML model classifies anomalies and filters noise in real time.
- Measurement overlay translates pixel distances into real-world units for reporting.
Python
OpenCV
Machine Learning
Image Processing
Result
Reduced manual inspection time while increasing detection accuracy for dents and cracks across automotive and manufacturing subsystems.