Cross-Disciplinary Image-Based Approaches for Heritage Analysis and Preservation

Authors

Mohammad Ibrahim Abu-Haifa, Florida International UniversityFollow

Document Type

Dissertation

Degree

Doctor of Philosophy (PhD)

Major/Program

Civil Engineering

First Advisor's Name

Seung Jae Lee

First Advisor's Committee Title

Committee chair

Second Advisor's Name

Armin Mehrabi

Second Advisor's Committee Title

committee member

Third Advisor's Name

David Garber

Third Advisor's Committee Title

committee member

Fourth Advisor's Name

Nipesh Pradhananga

Fourth Advisor's Committee Title

committee member

Keywords

Masonry structure, Image-based analysis, Polygon approximation, Discrete element method, Impulse-based dynamic simulation, Historic paintings, craquelure patterns analysis, Image segmentation and processing, Heritage preservation

Date of Defense

6-26-2023

Abstract

Preserving our cultural heritage requires ensuring the protection and maintenance of its tangible aspects such as historic structures and artworks. Therefore, their comprehensive analysis is a critical step towards their preservation. This study focuses on two key areas: unreinforced masonry structures and the craquelure patterns found in historic paintings by presenting two novel image-based analysis frameworks with the goal of developing more effective methods for preservation and restoration. The first framework utilizes image processing techniques, impulse based dynamics, and discrete element method, resulting in a streamlined methodology for efficient vulnerability assessment of masonry structures. The framework involves using several techniques such as image segmentation, polygon approximation, and geometric analysis methods. The splitting and Douglas-Peuker polygon approximation methods are adopted to enhance computational efficiency of the analysis. Even if the splitting method slightly outperforms Douglas-Peuker method, results show that both methods can simplify brick shapes with a high-fidelity level. This framework is applied on a real single-wythe masonry structure, Stylite Tower in Jordan, to examine its applicability, by investigating the impact of an earthquake of a magnitude 7 vii Richter on the structure, which can provide critical domain knowledge that can be useful for disaster preparedness.

This dissertation also highlights a phenotypic trait of historic paintings. This approach uses image processing techniques to convert the craquelure image into binary image, where the islands are presented in white pixels and the cracks in black. Then, geometric analysis is conducted by converting each island into a polygon and finding its area and perimeter. The obtained data are plotted in a log-log scale, where the x-axis represents perimeter/area and the y-axis represents the area, and the probability distribution of the data are determined in terms of bivariate ellipses. Results show that image processing helps in investigating dating and provenance of craquelure patterns, as they have a phenotypic trait, which can aid historians preserving ancient artworks. Also, it is noted that when historic paintings are from the same origin and time period, they are more likely to exhibit similar craquelure behavior.

Overall, the research makes significant contributions to the fields of engineering and art preservation.

Identifier

FIDC011109

ORCID

https://orcid.org/0009-0009-3554-8774

Previously Published In

1. Abu-Haifa, M. and Lee, S.J., 2022. Image-based modeling-to-simulation of masonry walls. Journal of Architectural Engineering, 28(4), p.06022001. https://doi.org/10.1061/(ASCE)AE.1943-5568.0000569
   
   
2. Abu-Haifa, M. and Lee, S.J., 2023. Image-Based 3D Modeling-to-Simulation of Single-wythe Masonry Structure Via Reverse Descriptive Geometry. Journal of Building Engineering, 76(10), p.107125. https://doi.org/10.1016/j.jobe.2023.107125

Recommended Citation

Abu-Haifa, Mohammad Ibrahim, "Cross-Disciplinary Image-Based Approaches for Heritage Analysis and Preservation" (2023). FIU Electronic Theses and Dissertations. 5326.
https://digitalcommons.fiu.edu/etd/5326