Building identification in aerial imagery using deep learning

Simple item page
dc.contributor.advisorGrobler, Trienkoen_ZA
dc.contributor.authorNakiranda, Proscoviaen_ZA
dc.contributor.otherStellenbosch University. Faculty of Science. Dept. of Computer Science.en_ZA
dc.date.accessioned2024-02-09T12:42:07Z
dc.date.accessioned2024-04-27T00:12:21Z
dc.date.available2024-02-09T12:42:07Z
dc.date.available2024-04-27T00:12:21Z
dc.date.issued2024-03
dc.descriptionThesis (MSc)--Stellenbosch University, 2024.en_ZA
dc.description.abstractENGLISH ABSTRACT: Advancements in the field of remote sensing have facilitated the effortless re- trieval of information about any location on Earth at any given time. This has resulted in a variety of developments, including the identification of economic activities taking place in a particular area. The task of identifying build- ings is one significant application of remote sensing imagery as it is crucial for assessing resource distribution, especially in low-resource or data-scarce areas. Advances in machine learning and computation resources allow for au- tomatic analysis of collected remote sensing data, eliminating the need for human intervention. The task of building identification falls under computer vision and is an example of a task that can be automated. Several machine learning architectures (models) have been proposed for building identification. However, choosing the appropriate one can be challenging due to limitations such as difficulty in accurately identifying boundary or near boundary pixels, resource requirements, and overall model accuracy. Therefore, conducting a comparative study is necessary to evaluate the performance of the building identification models. In this thesis, we carry out a comparative study of four state-of-the-art models used for the building identification task. We eval- uate their performance both qualitatively and quantitatively. Furthermore, we investigate the effect of multitask learning on the models’ performance in building identification. The thesis concludes by providing our research find- ings and outlining prospective future research avenues. Moreover, it provides a thorough overview of the fundamental theory underpinning remote sensing and machine learning.en_ZA
dc.description.abstractAFRIKAANSE OPSOMMING: Vooruitgang in die veld van afstandaardobservasie het dit moontlik gemaak om moeiteloos inligting van enige plek op aarde te bekom. Dit het ook tot verskeie ontwikkelings gelei, insluitende die vermoë om die ekonomiese aktiwiteite wat plaasvind in ’n gegewe area te identifiseer. Die taak om geboue te idetifiseer is een van die mees belangrikste toepassings in afstandaardobservasie omdat dit krities is vir effektiewe hulpbronbeplanning. Dit is veral belangrik vir ge- biede waar daar beperkte hulpbronne is. Vooruitgang in masjienleer en harde- ware maak dit moontlik om sekere take te automatiseer. Die taak van gebou- identifiseering is so ’n toepassing wat onder die veld van rekenaarvisie val. Ver- skeie masjienleer argitekture (modelle) word gebruik vir hierdie taak. Om die beste model te kies vir ’n spesifieke gebruiksgeval is nie altyd maklik nie, omdat verskillende dinge in ag geneem moet word wanneer so ’n besluit gemaak moet word. Dinge soos hoe akkuraat kan randpieksels onderskei word, beskikbare verwerkingskrag en die alghele akkuraatheid van modele moet in ag geneem word. Dit is hoekom vergelykingstudies van gebou-identifiseeringsargitekture van kardinale belang is. In hierdie tesis, word ’n vergelykingstudie tussen vier algeheel gebruikte gebou-identifiseeringsargitekture uitgevoer. Hierdie studie is beide kwantitatief en kwalitatief. Verder word daar ook ondersoek watter effek veeltaak afrigting het op gebou-identifiseeringsargitekture. Die tesis sluit af met die bevindinge van die studie en daar word voorstelle gemaak oor wat in die toekoms gedoen kan word. Die tesis gee ook ’n goeie oorsig van die volgende velde: masjienleer en afstandaardobservasie.af_ZA
dc.description.versionMastersen_ZA
dc.format.extentxvi, 109 pages : illustrations (some color)en_ZA
dc.identifier.urihttps://scholar.sun.ac.za/handle/10019.1/130619
dc.language.isoen_ZAen_ZA
dc.language.isoen_ZAen_ZA
dc.publisherStellenbosch : Stellenbosch Universityen_ZA
dc.rights.holderStellenbosch Universityen_ZA
dc.subject.lcshImage segmentationen_ZA
dc.subject.lcshRemote-sensing images -- Data processingen_ZA
dc.subject.lcshImage processing -- Digital techniquesen_ZA
dc.subject.lcshDeep learning (Machine learning) -- Technological innovationsen_ZA
dc.subject.lcshNeural networks (Computer science)en_ZA
dc.subject.lcshAerial photography -- Computer network resourcesen_ZA
dc.subject.nameUCTDen_ZA
dc.titleBuilding identification in aerial imagery using deep learningen_ZA
dc.typeThesisen_ZA
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
nakiranda_building_2024.pdf
Size:
8.54 MB
Format:
Adobe Portable Document Format
Description:
Download
Collections
Masters Degrees (Computer Science)