東華大學圖書館 |

Improving Deep Learning-Based Facade Visual Inspection: A Data Quality Perspective.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Improving Deep Learning-Based Facade Visual Inspection: A Data Quality Perspective./
Author:	Jingjing, Guo.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2021,
Description:	335 p.
Notes:	Source: Dissertations Abstracts International, Volume: 84-04, Section: B.
Contained By:	Dissertations Abstracts International84-04B.
Subject:	Deep learning. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29352297
ISBN:	9798352681305

Improving Deep Learning-Based Facade Visual Inspection: A Data Quality Perspective.
Jingjing, Guo.

Improving Deep Learning-Based Facade Visual Inspection: A Data Quality Perspective. - Ann Arbor : ProQuest Dissertations & Theses, 2021 - 335 p.

Source: Dissertations Abstracts International, Volume: 84-04, Section: B.

Thesis (Ph.D.)--National University of Singapore (Singapore), 2021.

This item must not be sold to any third party vendors.

A building facade is an external structure to support and protect a building. Defects occurring on facade can reduce the service life of the whole building. If timely inspection and maintenance are not conducted, facade defects may cause great damages to the surroundings. Therefore, facade inspection has drawn great attention in both industry and academia.Visual inspection that assesses the facade condition is the first and a key step for periodic facade maintenance. However, traditional approaches to achieving the facade visual inspection are criticized as risky, laborious, time-consuming, and expensive. To avoid the problems in the traditional approaches, there are increasing efforts exploring to use mobile devices (i.e., unmanned aerial vehicles and robots) and artificial intelligence techniques to inspect the facade condition. Particularly, deep learning algorithms have attracted considerable research interests in recent years for structural health monitoring. Although attempts have been made to achieve defects identification with different deep learning models for various structures, many problems still existed that weaken the performance of deep learning-based facade visual inspection. Especially, the problems caused by data quality were not well addressed in previous studies.The main objective of this thesis was to improve the performance of deep learning-based facade visual inspection. In this thesis, the performance was considered from the aspects of reliability and efficiency. To achieve this objective, this thesis designed a research methodology based on the theory of total data quality management. The methodology included four phases: definition, assessment, analysis, and improvement.In the first phase, the requirement of data quality was defined. Meanwhile, the critical procedures, activities, and human factors were extracted using a Delphi study. The results of the Delphi study revealed that the procedures of data pre-processing and model construction have greater influences on the uncertainty of efficiency and the uncertainty of reliability. Next, four data quality problems, including imbalanced distribution, incomplete information, ineffective labels, and inconsistent labels, were identified in the second phase of assessment. Then, in the third phase of analysis, the potential solutions to the four data quality problems were analyzed based on the idea of managing the existent data without burdening humans. In the last phase of improvement, a research framework consisting of the corresponding solutions was proposed. The research framework was unfolded through three procedures: data selection, data annotation, and model training. For each procedure, criteria were designed to assess the data quality, and solutions were developed enabling the target stage to focus on "better" data. For data selection, a semi-supervised learning solution and an active learning solution were designed by integrating the degree of uncertainty and the degree of representativeness to utilize the unlabeled data and to enrich the labeled data. For data annotation, annotation rules for categorization, localization, and segmentation were defined in accordance with the assessment standards for facade inspection. For model training, a meta learning method was applied for solving the problem of imbalanced distribution while a rule-based deep learning method was designed to regulate the learning direction. These two methods were combined with a criterion of the similarity to the ground-truth data to provide higher weights to the highquality data. The experiment results demonstrated that the proposed solutions successfully improved the accuracy and stability of the detection of various facade defects. Besides, the detection results obtained by the proposed solutions provided more effective outcomes for condition evaluation. Meanwhile, the time and cost were saved in general perspective because no extra labor works were expended. Therefore, the reliability and the efficiency of deep learning-based facade visual inspection were effectively improved by the proposed research framework.

ISBN: 9798352681305Subjects--Topical Terms:

3554982
Deep learning.

Improving Deep Learning-Based Facade Visual Inspection: A Data Quality Perspective.
LDR:05323nmm a2200385 4500 001 2393591
005 20240414211445.5
006 m o d
007 cr#unu||||||||
008 251215s2021 ||||||||||||||||| ||eng d
020 $a 9798352681305
035 $a (MiAaPQ)AAI29352297
035 $a (MiAaPQ)USingapore218220
035 $a AAI29352297
040 $a MiAaPQ $c MiAaPQ
100 1 $a Jingjing, Guo. $3 3763065
245 1 0 $a Improving Deep Learning-Based Facade Visual Inspection: A Data Quality Perspective.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2021
300 $a 335 p.
500 $a Source: Dissertations Abstracts International, Volume: 84-04, Section: B.
500 $a Advisor: Qian, Wang.
502 $a Thesis (Ph.D.)--National University of Singapore (Singapore), 2021.
506 $a This item must not be sold to any third party vendors.
520 $a A building facade is an external structure to support and protect a building. Defects occurring on facade can reduce the service life of the whole building. If timely inspection and maintenance are not conducted, facade defects may cause great damages to the surroundings. Therefore, facade inspection has drawn great attention in both industry and academia.Visual inspection that assesses the facade condition is the first and a key step for periodic facade maintenance. However, traditional approaches to achieving the facade visual inspection are criticized as risky, laborious, time-consuming, and expensive. To avoid the problems in the traditional approaches, there are increasing efforts exploring to use mobile devices (i.e., unmanned aerial vehicles and robots) and artificial intelligence techniques to inspect the facade condition. Particularly, deep learning algorithms have attracted considerable research interests in recent years for structural health monitoring. Although attempts have been made to achieve defects identification with different deep learning models for various structures, many problems still existed that weaken the performance of deep learning-based facade visual inspection. Especially, the problems caused by data quality were not well addressed in previous studies.The main objective of this thesis was to improve the performance of deep learning-based facade visual inspection. In this thesis, the performance was considered from the aspects of reliability and efficiency. To achieve this objective, this thesis designed a research methodology based on the theory of total data quality management. The methodology included four phases: definition, assessment, analysis, and improvement.In the first phase, the requirement of data quality was defined. Meanwhile, the critical procedures, activities, and human factors were extracted using a Delphi study. The results of the Delphi study revealed that the procedures of data pre-processing and model construction have greater influences on the uncertainty of efficiency and the uncertainty of reliability. Next, four data quality problems, including imbalanced distribution, incomplete information, ineffective labels, and inconsistent labels, were identified in the second phase of assessment. Then, in the third phase of analysis, the potential solutions to the four data quality problems were analyzed based on the idea of managing the existent data without burdening humans. In the last phase of improvement, a research framework consisting of the corresponding solutions was proposed. The research framework was unfolded through three procedures: data selection, data annotation, and model training. For each procedure, criteria were designed to assess the data quality, and solutions were developed enabling the target stage to focus on "better" data. For data selection, a semi-supervised learning solution and an active learning solution were designed by integrating the degree of uncertainty and the degree of representativeness to utilize the unlabeled data and to enrich the labeled data. For data annotation, annotation rules for categorization, localization, and segmentation were defined in accordance with the assessment standards for facade inspection. For model training, a meta learning method was applied for solving the problem of imbalanced distribution while a rule-based deep learning method was designed to regulate the learning direction. These two methods were combined with a criterion of the similarity to the ground-truth data to provide higher weights to the highquality data. The experiment results demonstrated that the proposed solutions successfully improved the accuracy and stability of the detection of various facade defects. Besides, the detection results obtained by the proposed solutions provided more effective outcomes for condition evaluation. Meanwhile, the time and cost were saved in general perspective because no extra labor works were expended. Therefore, the reliability and the efficiency of deep learning-based facade visual inspection were effectively improved by the proposed research framework.
590 $a School code: 1883.
650 4 $a Deep learning. $3 3554982
650 4 $a Defects. $3 3682384
650 4 $a Computer science. $3 523869
650 4 $a Unmanned aerial vehicles. $3 3560267
650 4 $a Boxes. $3 3564918
650 4 $a Localization. $3 3560711
650 4 $a Active learning. $3 527777
650 4 $a Neural networks. $3 677449
650 4 $a Classification. $3 595585
650 4 $a Design. $3 518875
650 4 $a Algorithms. $3 536374
650 4 $a Annotations. $3 3561780
650 4 $a Aerospace engineering. $3 1002622
650 4 $a Robotics. $3 519753
650 4 $a Accuracy. $3 3559958
650 4 $a Datasets. $3 3541416
650 4 $a Questionnaires. $3 529568
650 4 $a Efficiency. $3 753744
650 4 $a Research methodology. $3 3559994
650 4 $a Experiments. $3 525909
650 4 $a Data collection. $3 3561708
690 $a 0389
690 $a 0984
690 $a 0800
690 $a 0538
690 $a 0454
690 $a 0771
710 2 $a National University of Singapore (Singapore). $3 3352228
773 0 $t Dissertations Abstracts International $g 84-04B.
790 $a 1883
791 $a Ph.D.
792 $a 2021
793 $a English
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29352297