| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

Show document Help

Title:Hiperspektralno slikanje, difuzni modeli in modeli transformer za izboljšano detektiranje neeksplodiranih ubojnih sredstev : magistrsko delo
Authors:ID Vilec Letonja, Vid (Author)
ID Potočnik, Božidar (Mentor) More about this mentor... New window
ID Bajić, Milan (Comentor)
Files:.pdf MAG_Vilec_Letonja_Vid_2026.pdf (3,12 MB)
MD5: 3BA138602616298A4E587E4DB1F8EF3C
 
Language:Slovenian
Work type:Master's thesis/paper
Typology:2.09 - Master's Thesis
Organization:FERI - Faculty of Electrical Engineering and Computer Science
Abstract:V magistrskem delu raziskujemo uporabo hiperspektralnega slikanja v kombinaciji z difuznimi modeli in modeli Transformer za detekcijo neeksplodiranih ubojnih sredstev (UXO). Dimenzionalnost hiperspektralnih podatkov zmanjšamo z metodo PCA ter tako pripravljene podatke uporabimo kot vhod v različne arhitekture globokega učenja, med drugim Vision Transformer, SegFormer, Swin Transformer, TransUNet in difuzne modele. Obravnavamo tako binarno kot semantično segmentacijo objektov UXO. Uspešnost modelov ovrednotimo na javno dostopni zbirki 134 hiperspektralnih slik, zajetih s kamero Specim IQ, pri čemer uporabimo v raziskovalni skupnosti uveljavljene metrike. Najboljše rezultate doseže doučeni model SegFormer, kar potrjuje učinkovitost prenosa znanja iz segmentacije RGB-slik na segmentacijo hiperspektralnih slik. Difuzni model dosega obetavne rezultate, vendar je zaradi dolgega časa inferenc manj primeren za realnočasovno detekcijo.
Keywords:hiperspektralne slike, neeksplodirana ubojna sredstva, semantična segmentacija, modeli Transformer, difuzni modeli
Place of publishing:Maribor
Place of performance:Maribor
Publisher:[V. Vilec Letonja]
Year of publishing:2026
Number of pages:1 spletni vir (1 datoteka PDF (IX, 51 str.))
PID:20.500.12556/DKUM-97703 New window
UDC:004.932:004.85(043.2)
COBISS.SI-ID:278525955 New window
Publication date in DKUM:08.05.2026
Views:201
Downloads:21
Metadata:XML DC-XML DC-RDF
Categories:KTFMB - FERI
:
Copy citation
  
Average score:(0 votes)
Your score:Voting is allowed only for logged in users.
Share:Bookmark and Share



Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Licences

License:CC BY-NC 4.0, Creative Commons Attribution-NonCommercial 4.0 International
Link:http://creativecommons.org/licenses/by-nc/4.0/
Description:A creative commons license that bans commercial use, but the users don’t have to license their derivative works on the same terms.
Licensing start date:03.04.2026

Secondary language

Language:English
Title:Hyperspectral imaging, diffusion and transformer models for improved detection of unexploded ordnance
Abstract:In this master’s thesis, we investigate the use of hyperspectral imaging in combination with diffusion models and Transformer-based models for the detection of unexploded ordnance (UXO). The dimensionality of hyperspectral data is reduced using PCA, and the processed data is then used as input for various deep learning architectures, including Vision Transformer, SegFormer, Swin Transformer, TransUNet, and diffusion models. Both binary and semantic segmentation of UXO objects are addressed. Model performance is evaluated on a publicly available dataset of 134 hyperspectral images acquired with a Specim IQ camera, using metrics commonly adopted in the research community. The best results are achieved by a fine-tuned SegFormer model, confirming the effectiveness of knowledge transfer from RGB image segmentation to hyperspectral image segmentation. The diffusion model shows promising results, but due to long inference times, it is less suitable for real-time applications.
Keywords:hyperspectral images, unexploded ordnance, semantic segmentation, Transformers, diffusion models


Comments

Leave comment

You must log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back
Logos of partners University of Maribor University of Ljubljana University of Primorska University of Nova Gorica