arthurflor23/surface-crack-detection

Deep Learning Model for Crack Detection and Segmentation

Surface Crack Detection and Segmentation

Digital Image Processing class project, Master's degree, UPE - POLI (2018.2).

This project aims to detect and segment surface cracks in images using Digital Image Processing and deep learning techniques. The U-Net model was used for both crack detection and segmentation processes. For more details on the U-Net model, see the paper and the repository.

Results

Cracked Tile Detection

Original	Predicted	Overlay

Cracked Concrete Detection

Original	Predicted	Overlay

Citation

If this project has been helpful for your research, please consider citing the following paper:

@article{10.1590/s1678-86212021000100498,
   title     = {{Processamento digital de imagens para detec\~A\S\~A\poundso autom\~A!`tica de fissuras em revestimentos cer\~A\textcentmicos de edif\~A\-cios}},
   journal   = {{Ambiente Constru\~A\-do}},
   author    = {Ruiz, Ramiro Daniel Ballesteros AND Lordsleem Junior, Alberto Casado AND Neto, Arthur Flor de Sousa AND Fernandes, Bruno Jos\~A\copyright Torres},
   pages     = {139 - 147},
   volume    = {21},
   month     = {01},
   year      = {2021},
   publisher = {scielo},
   isbn      = {1678-8621},
   url       = {http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1678-86212021000100139&nrm=iso},
   doi       = {10.1590/s1678-86212021000100498},
}