FIBER SEGMENTATION IN CRACK REGIONS OF STEEL FIBER REINFORCED CONCRETE USING PRINCIPAL CURVATURE
Keywords:fiber reinforced concrete, fiber segmentation, image analysis, X-ray micro-computed tomography
AbstractThis paper tackles the non-trivial image-processing task to segment hook-ended fibers in three-dimensional images. For this purpose, a novel segmentation method is presented that relies on the following observation: For a single fiber the configurations of principal curvatures that can occur on its surface are limited. Deviations from these configurations indicate potential overlaps of fibers. The method that was developed based on this observation is used to separate several simulated clusters of touching fibers as a proof-of-concept. Further, it is applied to two images of cracked steel fiber reinforced concrete specimens arising from a 4-point bending test. The method's performance is compared to manual separation. Overall, we can state that the proposed method yields satisfying results when data meets the following criteria: Low fiber volume density, circular fiber cross section and sufficient spatial resolution of fiber-fiber contacts.
How to Cite
Kronenberger, M., Schladitz, K., Hamann, B., & Hagen, H. (2018). FIBER SEGMENTATION IN CRACK REGIONS OF STEEL FIBER REINFORCED CONCRETE USING PRINCIPAL CURVATURE. Image Analysis and Stereology, 37(2), 127–137. https://doi.org/10.5566/ias.1914
Original Research Paper
Copyright (c) 2018 Image Analysis & Stereology
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.