DeepLabV3+ Based Mask R-CNN for Crack Detection and Segmentation in Concrete Structures

In order to solve the problem of concrete structure crack detection and segmentation and improve the efficiency of detection and segmentation, this paper proposes a crack detection and segmentation method for concrete structure based on DeepLabV3+ and Mask R-CNN algorithm. Firstly, a crack detection and segmentation scheme is designed by analysing the crack detection and segmentation problem of concrete structure. Secondly, a crack detection method based on Mask R-CNN algorithm is proposed for the crack detection problem of concrete structure. Then, a crack segmentation method based on DeepLabV3+ algorithm is proposed for the crack segmentation problem of concrete structure. Finally, bridge crack image data is used for the crack detection and segmentation of concrete structure. Finally, the concrete structure crack detection and segmentation method is validated and analysed using bridge crack image data. The results show that the Mask R-CNN model has better performance in the localisation and identification of cracks, and the DeepLabV3+ model has higher accuracy and contour extraction integrity in solving the crack segmentation problem.

@article{Liu2025,
  title     = {DeepLabV3+ Based Mask R-CNN for Crack Detection and Segmentation in Concrete Structures},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {16},
  number    = {1},
  year      = {2025},
  publisher = {The Science and Information Organization},
  author    = {Yuewei Liu},
  doi       = {10.14569/IJACSA.2025.0160142},
  url       = {https://doi.org/10.14569/IJACSA.2025.0160142}
}