Richard Olamide Olanite
Back to all work

Measuring concrete crack width from photographs

Design codes judge cracks by width (Eurocode 2 recommends about 0.3 mm). So instead of asking only whether a model finds cracks, I asked whether it can measure them well enough for an inspector.

0.74Dice on held-out photos (95% CI 0.71–0.77)
0.77Dice on an unseen concrete dataset
79%test photos leaking into training in the official split
12 pxminimum crack width for ±25% accuracy

The problem

Crack segmentation papers usually report overlap scores. Inspectors need widths in millimetres, and a model that draws every crack a few pixels too wide can score well while being useless for that.

What I did

  1. Traced the 9,159 images of CrackSeg9k back to about 1,635 original photographs, and found that 79% of the official test photos have crops in the training set. Re-split by original photograph.
  2. Trained a U-Net with a ResNet-34 encoder, and tested on held-out photos and on three datasets never used in training (concrete, asphalt, tiles).
  3. Converted predicted outlines into crack length and width, compared against the labels, and simulated camera distance by resizing to find where width becomes reliable.

Results

Left: predicted against true crack width at six image scales; thin cracks come out too wide. Right: the width error stays within ±25% only above about 12 px.
Left: predicted against true crack width at six image scales; thin cracks come out too wide. Right: the width error stays within ±25% only above about 12 px.
Predicted vs labelled crack width.
Predicted vs labelled crack width.

Limitations

  • Thin cracks are always drawn 2–4 px too wide, so width is reliable only above about 12 px: about 7 cm from the wall with a 12 MP phone camera, for a 0.3 mm crack.
  • The model does not transfer to new materials: Dice 0.60 on asphalt and 0.29 on ceramic tiles.
  • Width 'truth' is the dataset labels, whose style varies by source. Field validation with hand-measured cracks is in progress.