東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Exploration of Multi-domain Data Fus...

Arnous, Ferris,

Linked to FindBook

Google Book

Amazon

博客來

Exploration of Multi-domain Data Fusion and Deep Learning Algorithms for Radar Imagery Classification /

Record Type:	Electronic resources : Monograph/item
Title/Author:	Exploration of Multi-domain Data Fusion and Deep Learning Algorithms for Radar Imagery Classification // Ferris Arnous.
Author:	Arnous, Ferris,
Description:	1 electronic resource (162 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 86-05, Section: B.
Contained By:	Dissertations Abstracts International86-05B.
Subject:	Remote sensing. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31631242
ISBN:	9798346387886

Exploration of Multi-domain Data Fusion and Deep Learning Algorithms for Radar Imagery Classification /
Arnous, Ferris,

Exploration of Multi-domain Data Fusion and Deep Learning Algorithms for Radar Imagery Classification /Ferris Arnous. - 1 electronic resource (162 pages)

Source: Dissertations Abstracts International, Volume: 86-05, Section: B.

Synthetic aperture radar (SAR) imaging has applications ranging from commercial non-destructive evaluation to defense-based remote sensing. Electromagnetic (EM) imaging with SAR however remains highly sensitive due to backscatter mechanisms that arise from resolving objects from complex environments. These mechanisms often create larger variations of features within the radar imagery, making tasks such as radar image classification with machine learning difficult. Thus, the need for enhancing robustness of deep learning algorithms to be able to identify newly unseen backscatter profiles that arise from these sensitivities, remains a challenge.This dissertation explores the use of sparse data augmentation and feature data fusion in the design and assessment of deep machine-learning algorithms for radar imagery classification. The motivation behind these implementations is to increase the ability of the algorithms to identify backscatter patterns not previously learned while limiting the number of examples the algorithms need to train on to do so. Thus, the designed approaches needed to be able to simultaneously adapt to a wide range of representations of objects and environments using few or single inputs. Highly flexible deep learning algorithms using convolutional neural networks (CNNs), long short-term memory networks (LSTMs) and visual transformer networks (ViTs) were developed to explore this capability. Inputs were sourced from fundamental and simulated dual modality and polarization radar images, multi-temporal radar image instances and spectral-image pairs. The results demonstrate an approach to combine advantages from multiple deep learning feature extraction methods while leveraging wider representations of the data from multiple augmented inputs. These approaches proved to enhance robustness on testing set electromagnetic (EM) image patterns that varied from training data along three simultaneous variations in representation, all while using fewer training instances.

English

ISBN: 9798346387886Subjects--Topical Terms:

535394
Remote sensing.

Exploration of Multi-domain Data Fusion and Deep Learning Algorithms for Radar Imagery Classification /
LDR:03511nmm a22004093i 4500 001 2435934
005 20251014050544.5
006 m o d
007 cr|nu||||||||
008 261011s2024 miu||||||m |||||||eng d
020 $a 9798346387886
035 $a (MiAaPQD)AAI31631242
035 $a (MiAaPQD)PennState25396fia5095
035 $a AAI31631242
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Arnous, Ferris, $e author. $3 3826423
245 1 0 $a Exploration of Multi-domain Data Fusion and Deep Learning Algorithms for Radar Imagery Classification / $c Ferris Arnous.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (162 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 86-05, Section: B.
500 $a Advisors: Lanagan, Michael T.; Narayanan, Ram M. Committee members: Segall, Albert; Kumara, Soundar; Shokouhi, Parisa; Peco, Christian; Kane, Timothly J.
502 $b Ph.D. $c The Pennsylvania State University $d 2024.
520 $a Synthetic aperture radar (SAR) imaging has applications ranging from commercial non-destructive evaluation to defense-based remote sensing. Electromagnetic (EM) imaging with SAR however remains highly sensitive due to backscatter mechanisms that arise from resolving objects from complex environments. These mechanisms often create larger variations of features within the radar imagery, making tasks such as radar image classification with machine learning difficult. Thus, the need for enhancing robustness of deep learning algorithms to be able to identify newly unseen backscatter profiles that arise from these sensitivities, remains a challenge.This dissertation explores the use of sparse data augmentation and feature data fusion in the design and assessment of deep machine-learning algorithms for radar imagery classification. The motivation behind these implementations is to increase the ability of the algorithms to identify backscatter patterns not previously learned while limiting the number of examples the algorithms need to train on to do so. Thus, the designed approaches needed to be able to simultaneously adapt to a wide range of representations of objects and environments using few or single inputs. Highly flexible deep learning algorithms using convolutional neural networks (CNNs), long short-term memory networks (LSTMs) and visual transformer networks (ViTs) were developed to explore this capability. Inputs were sourced from fundamental and simulated dual modality and polarization radar images, multi-temporal radar image instances and spectral-image pairs. The results demonstrate an approach to combine advantages from multiple deep learning feature extraction methods while leveraging wider representations of the data from multiple augmented inputs. These approaches proved to enhance robustness on testing set electromagnetic (EM) image patterns that varied from training data along three simultaneous variations in representation, all while using fewer training instances.
546 $a English
590 $a School code: 0176
650 4 $a Remote sensing. $3 535394
650 4 $a Deep learning. $3 3554982
650 4 $a Fourier transforms. $3 3545926
650 4 $a Decision making. $3 517204
650 4 $a Signal processing. $3 533904
650 4 $a Neural networks. $3 677449
650 4 $a Radar. $3 594027
650 4 $a Visualization. $3 586179
650 4 $a Geometry. $3 517251
650 4 $a Electrical engineering. $3 649834
650 4 $a Mathematics. $3 515831
690 $a 0799
690 $a 0800
690 $a 0544
690 $a 0405
710 2 $a The Pennsylvania State University. $e degree granting institution. $3 3765934
720 1 $a Lanagan, Michael T. $e degree supervisor.
720 1 $a Narayanan, Ram M. $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-05B.
790 $a 0176
791 $a Ph.D.
792 $a 2024
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31631242