Information and Diagnostics Systems

Visual Corrosion Assessment of Structural Steels Using Neural Networks

Oleh Venhryniuk (olehvenhryniuk@gmail.com)

Karpenko Physico-Mechanical Institute of NAS of Ukraine

https://doi.org/10.15407/msse2023.224

PDF DocumentPDF

Keywords

steel structures, visual corrosion assessment, machine learning, neural networks, DeepLabV3+, Yolov8x

Abstract

Comparison of different semantic segmentation models for visual corrosion assessment of structural steels. The work proposes results of applying Yolov8x and compares results with DeepLabv3+[3] for corrosion segmentation task.

References

  1. Witcher T.R. From disaster to prevention: The silver bridge // Civil engineering. – 2018. – 87. – P. 42-43. https://doi.org/10.1061/ciegag.0001238
  2. Broughton E. The Bhopal disaster and its aftermath: a review // Environ. Health. – 2005. – 4, № 1. -No. 15882472. https://doi.org/10.1186/1476-069X-4-6
  3. Bianchi E. and Hebdon M. Trained model for the semantic segmentation of corrosion condition states. – University Libraries: Virginia Tech., 2021. – Software.