Fully Convolutional Networks for Automatic Pavement Crack Segmentation

Uriel Escalona, Fernando Arce, Erik Zamora, Juan Humberto Sossa Azuela


Pavement cracks are an increasing threat to public safety. Automatic pavement crack segmentation remains a very challenging problem due to crack texture inhomogeneity, high outlier potential, large variability of topologies, and so on. Due to this, automatic pavement crack detection has captured the attention of the computer vision community, and a great quantity of algorithms for solving this task have been proposed. In this work, we study a U-Net network and two variants for automatic pavement crack detection. The main contributions of this research are: 1) two U-Net based network variations for automatic pavement crack detection, 2) a series of experiments to demonstrate that the proposed architectures outperform the state-of-the-art for automatic pavement crack detection using two public and well-known challenging datasets: CFD and AigleRN and 3) the code for this approach.


Automatic pavement crack detection, pavement cracks, fully convolutional neural networks

Full Text: PDF