Methods for segmenting cracks in 3d images of concrete: A comparison based on semi-synthetic images
作者：
Highlights：
• Comparison of eight methods for crack detection in 3d CT concrete images.
• Simulation of 3d crack images gives unbiased ground truth for evaluation.
• Parameter tuning with respect to different objectives.
• Machine learning methods (3d U-net and random forest) perform best among all methods.
• Hessian-based percolation performs best among classical methods.
摘要
•Comparison of eight methods for crack detection in 3d CT concrete images.•Simulation of 3d crack images gives unbiased ground truth for evaluation.•Parameter tuning with respect to different objectives.•Machine learning methods (3d U-net and random forest) perform best among all methods.•Hessian-based percolation performs best among classical methods.
论文关键词：Computed tomography,Fractional Brownian surface,3d segmentation,Crack detection,Machine learning,Deep learning
论文评审过程：Received 1 September 2021, Revised 19 April 2022, Accepted 24 April 2022, Available online 26 April 2022, Version of Record 13 May 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108747
