Crack Detection with YOLOv11

An automated computer-vision system for detecting and classifying cracks in civil infrastructure, built on YOLOv11. It targets the need for efficient, reliable, and scalable monitoring of structural integrity in bridges, roads, and buildings.

Approach

• Trained a YOLOv11 object-detection model to localize and classify cracks from infrastructure imagery.
• Built an end-to-end pipeline covering data preparation, training, evaluation, and inference on real crack datasets.
• Focused on a workflow that could scale to field inspection rather than a one-off lab demo.

Code and notebooks are on GitHub.