MARC ayrıntıları
| 000 -BAŞLIK |
|---|
| Sabit Uzunluktaki Kontrol Alanı |
03899nam a22003017a 4500 |
| 003 - KONTROL NUMARASI KİMLİĞİ |
|---|
| Kontrol Alanı |
KOHA |
| 005 - EN SON İŞLEM TARİHİ ve ZAMANI |
|---|
| Kontrol Alanı |
20231103100936.0 |
| 008 - SABİT UZUNLUKTAKİ VERİ ÖGELERİ - GENEL BİLGİ |
|---|
| Sabit Alan |
231103d2023 cy ||||| m||| 00| 0 eng d |
| 040 ## - KATALOGLAMA KAYNAĞI |
|---|
| Özgün Kataloglama Kurumu |
CY-NiCIU |
| Kataloglama Dili |
eng |
| Çeviri Kurumu |
CY-NiCIU |
| Açıklama Kuralları |
rda |
| 041 ## - DİL KODU |
|---|
| Metin ya da ses kaydının dil kodu |
eng |
| 090 ## - Yerel Tasnif No |
|---|
| tasnif no |
YL 3114 |
| Cutter no |
O46 2023 |
| 100 1# - KİŞİ ADI |
|---|
| Yazar Adı (Kişi adı) |
Omnondiagbe, David Agbonjague |
| 245 10 - ESER ADI BİLDİRİMİ |
|---|
| Başlık |
AUTOMATED CONCRETE CRACK DETECTION USING DEEP LEARNING TECHNIQUES / |
| Sorumluluk Bildirimi |
DAVID AGBONJAGUE OMONDIAGBE; SUPERVISOR: ASSOC. PROF. DR. TAMER TULGAR |
| 264 ## - PRODUCTION, PUBLICATION, DISTRIBUTION, MANUFACTURE, AND COPYRIGHT NOTICE |
|---|
| Date of production, publication, distribution, manufacture, or copyright notice |
2023 |
| 300 ## - FİZİKSEL TANIMLAMA |
|---|
| Sayfa, Cilt vb. |
xii, 116 sheets; |
| Boyutları |
31 cm. |
| Birlikteki Materyal |
1 CD-ROM |
| 336 ## - CONTENT TYPE |
|---|
| Source |
rdacontent |
| Content type term |
text |
| Content type code |
txt |
| 337 ## - MEDIA TYPE |
|---|
| Source |
rdamedia |
| Media type term |
unmediated |
| Media type code |
n |
| 338 ## - CARRIER TYPE |
|---|
| Source |
rdacarrier |
| Carrier type term |
volume |
| Carrier type code |
nc |
| 502 ## - TEZ NOTU |
|---|
| Tez Notu |
Thesis (MSc) - Cyprus International University. Institute of Graduate Studies and Research Information Systems Engineering Department |
| 504 ## - BİBLİYOGRAFİ NOTU |
|---|
| Bibliyografi Notu |
Includes bibliography (sheets 81-87) |
| 520 ## - ÖZET NOTU |
|---|
| Özet notu |
ABSTRACT<br/>Concrete is one of the most commonly used material in civil engineering and <br/>construction. It is mainly used in the construction of buildings, bridges, pavements etc. <br/>These concrete infrastructures are prone to cracks and these cracks are caused by <br/>various factors. Early detection of concrete cracks guarantees the safety of concrete <br/>infrastructures. The main focus of this work is to develop an automated deep learning <br/>model for detecting concrete cracks. This research uses the Concrete Crack Image <br/>Classification (CCIC) dataset having two classes. The dataset is balanced containing <br/>20,000 images with cracks (positive class) and 20,000 images without cracks (negative <br/>class).<br/>In this work, the effect of converting images to grayscale and fine tuning hyper <br/>parameters (batch size and Adam LR) was investigated by performing several <br/>experiments. Converting the images to grayscale resulted in simpler models with <br/>reduced parameters due to the one channel of grayscale images as compared to the <br/>three channels of RGB images. Two different architectures of CNN models, CNN 1 <br/>and CNN 2 were built and trained on RGB and grayscale images. The experimental <br/>results revealed that there is no much significant difference in the performance of the <br/>RGB and grayscale CNN models. This means that concrete crack detection relies more <br/>on structural patterns rather than color information. Although, converting the images <br/>to grayscale allows the model to focus more on learning the crack features.<br/>The models were further tested on an external data different from the development <br/>data. In this experiment, the CNN 2 model using grayscale images (batch size of 32 <br/>and Adam learning rate of 0.001) outperformed the other models and is chosen as the <br/>proposed model. The proposed model achieved a little change decrease in the <br/>difference in performance between the development data and external data.<br/>In addition, the proposed model’s generalization ability was further validated using 10 <br/>fold cross validation method. This was done to ensure that the model is not overfitting. <br/>The proposed model achieved a mean accuracy of 99.74%, mean precision of 99.83%, <br/>mean recall of 99.65%, mean F1-score of 99.74%, mean specificity of 99.83%, and <br/>mean AUC ROC of 0.9974. These performances were further compared with seven <br/>previous concrete crack detection studies that have used the CCIC dataset. From the <br/>comparative analysis, the proposed model performed better than the solutions of six <br/>of these studies in all the performance metrics compared.<br/>Keywords: Computer Vision, Concrete Crack Detection, Convolutional Neural <br/>Network, Deep Learning |
| 650 #0 - KONU BAŞLIĞI EK GİRİŞ - KONU TERİMİ |
|---|
| Konusal terim veya coğrafi ad |
Computer vision |
| Alt başlık biçimi |
Dissertations, Academic |
| 650 #0 - KONU BAŞLIĞI EK GİRİŞ - KONU TERİMİ |
|---|
| Konusal terim veya coğrafi ad |
Concrete |
| Alt başlık biçimi |
Dissertations, Academic |
| 650 #0 - KONU BAŞLIĞI EK GİRİŞ - KONU TERİMİ |
|---|
| Konusal terim veya coğrafi ad |
Neural networks (Computer science) |
| Alt başlık biçimi |
Dissertations, Academic |
| 650 #0 - KONU BAŞLIĞI EK GİRİŞ - KONU TERİMİ |
|---|
| Konusal terim veya coğrafi ad |
Deep learning (Machine learning) |
| Alt başlık biçimi |
Dissertations, Academic |
| 700 1# - EK GİRİŞ - KİŞİ ADI |
|---|
| Yazar Adı (Kişi adı) |
Tulgar, Tamer |
| İlişkili Terim |
supervisor |
| 942 ## - EK GİRİŞ ÖGELERİ (KOHA) |
|---|
| Sınıflama Kaynağı |
Dewey Onlu Sınıflama Sistemi |
| Materyal Türü |
Thesis |