Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields
Thesis (MEng)--Stellenbosch University, 2026.
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
Stellenbosch : Stellenbosch University
2026
|
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613772216008704 |
|---|---|
| access_status_str | Open Access |
| author | Phukubye, Noko Sekolo |
| author2 | Venter, Gerhardus |
| author_browse | Phukubye, Noko Sekolo Venter, Gerhardus |
| author_facet | Venter, Gerhardus Phukubye, Noko Sekolo |
| author_sort | Phukubye, Noko Sekolo |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Thesis (MEng)--Stellenbosch University, 2026. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/135942 |
| institution | Stellenbosch University (South Africa) |
| language | English |
| last_indexed | 2026-06-10T12:41:26.849Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2026 |
| publishDateRange | 2026 |
| publishDateSort | 2026 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/135942 Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields Phukubye, Noko Sekolo Venter, Gerhardus Neaves, Melody Stellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering. Thesis (MEng)--Stellenbosch University, 2026. Phukubye, N. S. 2026. Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/1aa568e6-0cd4-418f-a064-807060e3da98 Digital Image Correlation (DIC) is a non‑contact optical technique that relies on high‑contrast, stochastic speckle patterns to measure surface deformations. Recent research has focused on quantifying speckle pattern quality to improve measurement accuracy, leading to the development of several pattern quality metrics. This study builds on these efforts by using finite element (FE) methods to simulate surface deformations on various classes of speckle patterns with differing morphologies, providing ground‑truth displacement fields for DIC evaluation. Twelve distinct pattern classes were generated using two purpose‑built pattern generators for traditional speckles and checkerboard patterns, and a Perlin noise function. Additional classes were created by applying three texturing functions to the Perlin noise and by repeatedly blurring and sharpening the resulting images. Each class contained 600 greyscale, 8‑bit patterns. The patterns were evaluated using established quality metrics, including the Sum of Squared Subset Intensity Gradients (SSSIG), Mean Intensity Gradient (MIG), Mean Subset Fluctuation (MSF), Mean Intensity of Subset Differences (MIOSD), energy factor (𝐸𝑓), Shannon entropy, autocorrelation peak radius (𝑅𝑝𝑒𝑎𝑘), and a newly introduced metric, the Power Spectrum Area (PSA). Finite element simulations were performed in MSC Apex under constant strain conditions. The deformed images were analysed using Stellenbosch University’s SUN‑DIC software, which was integrated into a Python‑based workflow through application programming interfaces. The measured displacements were compared to the FE ground truth using the root mean squared error (RMSE). Correlations between the RMSE and the pattern metrics were examined to assess how spatial characteristics influence DIC accuracy. Strong correlations were observed between RMSE and the SSSIG, MIG, MIOSD, PSA, and 𝑅𝑝𝑒𝑎𝑘 within individual pattern classes, though these relationships weak-ened when data from all classes were combined. The best‑performing patterns, which exhibited the lowest RMSE values, originated from one of the Perlin noise classes and displayed feature sizes of approximately five pixels. Pattern optimisation was explored using support vector regression models trained on the pattern data and optimised with a trust‑constr algorithm, yielding improvements in one class. This study establishes a unified framework for generating, evaluating, and op-timising speckle patterns through FE‑based image deformation and DIC analysis. The PSA metric proved effective when evaluated across the full frequency spectrum and shows promise for more targeted applications involving selected frequency ranges. A multimetric approach is recommended for comprehensive pattern assessment. Masters 2026-04-15T14:05:44Z 2026-04-15T14:05:44Z 2026-03 Thesis https://scholar.sun.ac.za/handle/10019.1/135942 en Stellenbosch University 133 pages application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Phukubye, Noko Sekolo Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields |
| title | Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields |
| title_full | Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields |
| title_fullStr | Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields |
| title_full_unstemmed | Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields |
| title_short | Evaluating digital image correlation speckle pattern quality using finite element‑based displacement fields |
| title_sort | evaluating digital image correlation speckle pattern quality using finite element based displacement fields |
| url | https://scholar.sun.ac.za/handle/10019.1/135942 |
| work_keys_str_mv | AT phukubyenokosekolo evaluatingdigitalimagecorrelationspecklepatternqualityusingfiniteelementbaseddisplacementfields |