kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Detection of Wildfire Smoke Plumes Using YOLOv5
KTH, School of Electrical Engineering and Computer Science (EECS).
KTH, School of Electrical Engineering and Computer Science (EECS).
2024 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

This thesis investigates the potential of utilizing a YOLOv5model for detecting wildfire smoke plumes. The methodology involves severalessential steps. Initially, images containing smoke plumes are annotated, followed byaugmenting the dataset to improve its size and diversity. Various model parametersare then systematically explored and optimized through a manual search approach,aiming to maximize correct predictions while minimizing false predictions. The best-performing model in this thesis achieved a smoke plume detection rate of 75.6%.Furthermore, the effectiveness of using a model trained on difference images isexamined, demonstrating promising results with a detection rate of 72.6%. Futureresearch should explore alternative image processing methods, integrate additionaldata augmentation techniques, and investigate different YOLO architectures tofurther enhance detection performance in practical scenarios.

Abstract [sv]

Denna kandidatuppsats undersöker möjligheten attanvända en YOLOv5-modell för att upptäcka rökmoln från skogsbränder. Metodenomfattar flera väsentliga steg. Först annoteras bilder som innehåller rökmoln, följt avatt datasetet bildbehandlas för att utöka datasetets storlek. Olika modellparametrarutforskas och optimeras sedan systematiskt genom en manuell sökmetod, medmålet att maximera korrekta förutsägelser samtidigt som falska förutsägelserminimeras. Den bäst presterande modellen i denna undersökning uppnådde enupptäckningsgrad på 75.6% för rökmoln. Dessutom undersöks effektiviteten hos enmodell tränad på differensbilder, vilket visar lovande resultat med enupptäckningsgrad på 72.6%. Framtida forskning bör undersöka alternativa metoderför bildbehandling, integrera ytterligare tekniker för dataaugmentering och undersökaandra YOLO-arkitekturer för att ytterligare förbättra detektionsprestandan i praktiskascenarier.

Place, publisher, year, edition, pages
2024. , p. 485-493
Series
TRITA-EECS-EX ; 2024:177
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
URN: urn:nbn:se:kth:diva-359391OAI: oai:DiVA.org:kth-359391DiVA, id: diva2:1933162
Supervisors
Examiners
Projects
Kandidatexamensarbete Elektroteknik EECS 2024Available from: 2025-01-30 Created: 2025-01-30

Open Access in DiVA

fulltext(121150 kB)35 downloads
File information
File name FULLTEXT01.pdfFile size 121150 kBChecksum SHA-512
82342f1408fe2aae929f55e76f2a176a8521cf94c0fe100464225724a9b74ddff6f61a0fb1cf957b5d6400be30877b9d0aec28080253ae7c1e09793536e10217
Type fulltextMimetype application/pdf

By organisation
School of Electrical Engineering and Computer Science (EECS)
Electrical Engineering, Electronic Engineering, Information Engineering

Search outside of DiVA

GoogleGoogle Scholar
Total: 35 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 862 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf