Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Optimisation of digital radiography using multi-objective particle swarm optimisation
Thesis (PhD)--Stellenbosch University, 2025.
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Published: |
Stellenbosch : Stellenbosch University
2026
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867614030701527040 |
|---|---|
| access_status_str | Open Access |
| author | Nshimirimana, Robert Bellarmin |
| author2 | Engelbrecht, Andries |
| author_browse | Engelbrecht, Andries Nshimirimana, Robert Bellarmin |
| author_facet | Engelbrecht, Andries Nshimirimana, Robert Bellarmin |
| author_sort | Nshimirimana, Robert Bellarmin |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Thesis (PhD)--Stellenbosch University, 2025. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/134747 |
| institution | Stellenbosch University (South Africa) |
| last_indexed | 2026-06-10T12:45:33.890Z |
| 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:10019.1/134747 Optimisation of digital radiography using multi-objective particle swarm optimisation Nshimirimana, Robert Bellarmin Engelbrecht, Andries Abraham, Ajith Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering. Radiography, Medical -- Digital techniques Diagnostic imaging Optimisation . . . Thesis (PhD)--Stellenbosch University, 2025. Nshimirimana, R. B. 2025. Optimisation of digital radiography using multi-objective particle swarm optimisation. Unpublished doctoral dissertation. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/bd07cff8-caf0-484a-b9e4-f2b332f9eed3 ENGLISH ABSTRACT: Radiography is a technique that provides a two-dimensional (2-D) projection image that is extensively used for non-destructive investigations of materials and objects. The scientific community uses the technique to retrieve qualitative and quantitative information on the internal structure of a sample under investigation. The investigation is performed using the radiographic image produced by a radiography system of which parameters such as the degree of collimation, spectrum tailoring and the relative distances between source, object and detector can be adjusted. The process entails arranging parameters in such a way that a radiograph is produced that optimises those qualities desired for a given investigation. This can be done experimentally by changing parameters until a desired image quality (such as contrast or sharpness) is achieved. However, a manual approach to radiography process optimisation is an unsatisfactory time-consuming and labour-intensive process. The main focus of this thesis is on the automatic optimisation of a radiography system using a multi-objective particle swarm optimisation algorithm. The proposed optimisation technique is applied in a virtual environment, which represents a real radiography system. Other available radiography simulators that provide a virtual environment are unsuitable for a multi-objective particle swarm optimisation algorithm. This thesis proposes a radiography simulator tool that uses ray tracing and the Beer-Lambert Law to provide a suitable virtual environment for a multi-objective particle swarm optimisation algorithm. Furthermore, a novelty is added to the simulator to automatically evaluate the simulated radiograph in terms of image quality. Multi-objective particle swarm optimisation algorithms have been used in many real-world multi-objective optimisation problems, but performance of the algorithms is problem-dependent. The algorithms must be modified or fine-tuned to solve a particular real-world multi-objective optimisation problem. In this thesis, a multi-objective particle swarm optimisation algorithm suitable for the automation of radiography system optimisation is proposed. The new algorithm is a modified version of an existing multi-objective particle swarm optimisation algorithm. The new algorithm uses particle swarm optimisation (PSO) stability conditions, which guarantee that the swarm will reach a state of equilibrium. Thus, alleviating the necessity for fine-tuning these problem-dependent control parameters. The thesis closes by incorporating the new algorithm into a radiography simulator. Experimental analysis shows that the resulting radiography optimiser can solve real-world multi-objective radiography system optimisation problems, such as the design of a neutron collimator or the set-up parameters needed for a radiography scan. AFRIKAANSE OPSOMMING: Radiografie is 'n tegniek wat 'n twee-dimensionele (2-D) projeksiebeeld verskaf wat wyd gebruik word vir nie-vernietigende ondersoeke van materiale en voorwerpe. Die tegniek word deur die wetenskaplike gemeenskap aangewend om kwalitatiewe en kwantitatiewe inligting van die interne struktuur van 'n steekproef wat ondersoek word, te verkry. Die ondersoek word uitgevoer op die radiografiese beeld wat deur 'n radiografiestelsel geproduseer word waarvan parameters soos die graad van kollimasie, spektrumaanpassing en die relatiewe afstande tussen bron, voorwerp en detektor aangepas kan word. Die proses behels die rangskikking van parameters op so 'n wyse dat 'n radiograaf geproduseer word wat daardie eienskappe wat vir 'n gegewe ondersoek verlang word, optimaliseer. Dit kan eksperimenteel gedoen word deur parameters te verander totdat 'n gewenste beeldkwaliteit (soos byvoorbeeld kontras of skerpte) bereik word. 'n Handmatige benadering tot radiografieprosesoptimalisering is egter 'n onbevredigende tydrowende en arbeidsintensiewe proses. Hierdie tesis fokus op die outomatiese optimering van 'n radiografiestelsel, deur gebruik te maak van 'n multi-objektiewe deeltjielswerm optimeringsalgoritme. Die voorgestelde optimeringstegniek word toegepas in 'n virtuele omgewing wat 'n werklike radiografiestelsel verteenwoordig toegepas. Ander beskikbare radiografie-simulators wat 'n virtuele omgewing verskaf is nie geskik vir 'n multi-objektiewe deeltjie swerm optimeringsalgoritme nie. Hierdie tesis stel 'n radiografie-simulatorinstrument voor wat straalnasporing en die Beer-Lambert-wet gebruik om 'n geskikte virtuele omgewing vir 'n multi-objektiewe deeltjie swerm optimeringsalgoritme te verskaf. Verder word die simulator verbeter met ‘n byvoeging om die gesimuleerde radiografie outomaties in terme van beeldkwaliteit te evalueer. Multi-objektiewe deeltjie swerm optimeringsalgoritmes is in baie werklike multi-objektiewe optimeringsprobleme gebruik, maar prestasie van die algoritmes is probleemafhanklik. Die algoritmes moet aangepas of verfyn word om 'n spesifieke werklike multi-objektiewe optimeringsprobleem op te los. In hierdie tesis word 'n multi-objektiewe deeltjie swerm optimeringsalgoritme wat geskik is vir die outomatisering van radiografie stelsel optimerings voorgestel. Die nuwe algoritme is 'n gewysigde weergawe van in bestaande multi-objektiewe deeltjie swerm optimeringsalgoritme. Die nuwe algoritme gebruik deeltjie swerm optimering stabiliteitstoestande, wat waarborg dat die swerm 'n toestand van ewewig sal bereik. Hierdie verlig dus die noodsaaklikheid om die probleemafhanklike beheerparameters fyn in te stel. Die tesis sluit af deur die nuwe algoritme in 'n radiografiesimulator in te sluit. Eksperimentele analise toon dat die resulterende radiografie-optimaliseerder in staat is om werklike multi-objektiewe radiografiestelseloptimeringsprobleme op te los, soos die ontwerp van 'n neutronkollimator of die opstellingsparameters wat nodig is vir 'n radiografieskandering. Doctoral 2026-01-06T12:18:18Z 2026-01-06T12:18:18Z 2025-12 Thesis https://scholar.sun.ac.za/handle/10019.1/134747 Stellenbosch University xxxi, 238 pages : illustrations application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Radiography, Medical -- Digital techniques Diagnostic imaging Optimisation . . . Nshimirimana, Robert Bellarmin Optimisation of digital radiography using multi-objective particle swarm optimisation |
| title | Optimisation of digital radiography using multi-objective particle swarm optimisation |
| title_full | Optimisation of digital radiography using multi-objective particle swarm optimisation |
| title_fullStr | Optimisation of digital radiography using multi-objective particle swarm optimisation |
| title_full_unstemmed | Optimisation of digital radiography using multi-objective particle swarm optimisation |
| title_short | Optimisation of digital radiography using multi-objective particle swarm optimisation |
| title_sort | optimisation of digital radiography using multi objective particle swarm optimisation |
| topic | Radiography, Medical -- Digital techniques Diagnostic imaging Optimisation . . . |
| url | https://scholar.sun.ac.za/handle/10019.1/134747 |
| work_keys_str_mv | AT nshimirimanarobertbellarmin optimisationofdigitalradiographyusingmultiobjectiveparticleswarmoptimisation |