Automatic Crack Detection on Two-Dimensional Pavement Images: An Algorithm Based on Minimal Path Selection
Résumé
This paper proposes a new algorithm for automatic crack detection from pavement images. It heavily relies on the localization of minimal paths within each image, a path being a series of neighbouring pixels with low intensities. The originality of the approach is about the manner to select the set of minimal paths and the two post-processing steps introduced based on these minimal paths. Such an approach is a natural way to take account of both photometric and geometric characteristics of pavement images. The resulting crack detection method incorporates very few parameters. An intensive validation on both synthetic and real images is provided, with comparisons to existing methods.