An open-source digital image correlation software system.

Simple item page
dc.contributor.advisorVenter, Gen_ZA
dc.contributor.advisorConradie, PDFen_ZA
dc.contributor.authorBrisley, JEen_ZA
dc.contributor.otherStellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.en_ZA
dc.date.accessioned2024-02-26T11:28:39Zen_ZA
dc.date.accessioned2024-04-26T17:39:08Zen_ZA
dc.date.available2024-02-26T11:28:39Zen_ZA
dc.date.available2024-04-26T17:39:08Zen_ZA
dc.date.issued2024-02en_ZA
dc.descriptionThesis (MEng)--Stellenbosch University, 2024.en_ZA
dc.description.abstractENGLISH ABSTRACT:Digital Image Correlation (DIC) is a popular computer vision-based optical measurement technique typically used for material characterisation and load response tests. In this project, an open source software system in Python is developed for full-field DIC measurements. An overview of the numerical implementations underpinning local and global planar-DIC solvers for displacement measurements is provided. It is shown that the planar-DIC implementations accurately recover complex, sub-pixel level displacements, by testing the software on the synthetic image sets made available for the DIC challenge 2.0. The local-DIC implementation achieves metrological performance on par with the benchmark, while the global counterpart achieves metrological performance in the mid-range of the codes participating in the challenge. Strain filters implementing a Gaussian smoothing scheme and linear polynomial fit, respectively, to subsequently filter the displacement data are shown to achieve metrological performance on par with the benchmarks. Stereo-DIC combines the independent planar measurements from two cameras in a triangulation routine to estimate a depth component, and thus 3D fullfield displacement data. The stereo-DIC implementation is validated against displacements as set by a secondary measurement system, namely, high-precision optical translation stages. Additionally, the quantification of measurement uncertainty in stereo-DIC measurements is explored. A tensile test case study is performed to illustrate the influence of different user-defined software settings on the resolution characteristics of the full-field displacement and strain measurements obtained by stereo-DIC.en_ZA
dc.description.abstractAFRIKAANSE OPSOMMING: Digitale Beeld Korrelasie (DIC) is ’n gewilde rekenaarvisie-gebaseerde optiese meet tegniek wat tipies gebruik word vir materiaal karakterisering en belasting toetse. In hierdie projek is ’n oopbron DIC sagtewarestelsel ontwikkel in Python vir veplasing en vervorming metings met DIC. ’n Oorsig van die numeriese implementering vir plaaslike en globale 2DDIC metings word verskaf. Dit word aangetoon dat die 2D-DIC implementering komplekse, sub-piksel verplasings met hoë akkuraatheid kan meet, deur die sagteware te toets met sintetiese beelde wat beskikbaar gestel word in die DIC uitdaging 2.0. Die plaaslike-DIC implementering behaal meet akkuraatheid soortgelyk aan die top presterende kodes. Die globale implementering presteer op die middelvlak van al die deelnemers. Dit word aangetoon dat die vervormingsfilters wat: (i) ’n Gaussian gewig skema; en (ii) ’n lineêre polinoom passing gebruik, vervormings kan meet met akkuraatheid soortgelyk aan die toppresterende kodes. Stereo-DIC kombineer die onafhanklike 2D metings van twee kameras in ’n diepte-skatting proses, wat lei tot 3D verplasing metings. Die stereo-DIC implementering validasie behels ’n kwantitatiewe vergelyking met die verplasings soos bepaal deur hoë-akkuraatheid optiese skuiftafels. ’n Studie rakende die kwantifisering van onsekerheid in DIC metings is oorweeg. ’n Trektoets gevallestudie word uitgevoer om die invloed van verskillende sagetware instellings, bepaal deur die gebruiker, se invloed op die meetakkuraatheid te ondersoek.af_ZA
dc.description.versionMastersen_ZA
dc.format.extentxii, 102 pages : illustrations.en_ZA
dc.identifier.urihttps://scholar.sun.ac.za/handle/10019.1/130439en_ZA
dc.language.isoen_ZAen_ZA
dc.language.isoen_ZAen_ZA
dc.publisherStellenbosch : Stellenbosch Universityen_ZA
dc.rights.holderStellenbosch Universityen_ZA
dc.subject.lcshDigital image correlationen_ZA
dc.subject.lcshMetrologyen_ZA
dc.subject.lcshPython (Computer program language)en_ZA
dc.subject.lcshComputer visionen_ZA
dc.subject.lcshUCTDen_ZA
dc.titleAn open-source digital image correlation software system.en_ZA
dc.typeThesisen_ZA
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
brisley_open_2024.pdf
Size:
10.85 MB
Format:
Adobe Portable Document Format
Description:
Download
Collections
Masters Degrees (Mechanical and Mechatronic Engineering)