• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.221-235
• /
• 2022

Manual inspection of steel box girders on long span bridges is time-consuming and labor-intensive. The quality of inspection relies on the subjective judgements of the inspectors. This study proposes an automated approach to detect and segment cracks in high-resolution images. An end-to-end cascaded framework is proposed to first detect the existence of cracks using a deep convolutional neural network (CNN) and then segment the crack using a modified U-Net encoder-decoder architecture. A Naïve Bayes data fusion scheme is proposed to reduce the false positives and false negatives effectively. To generate the binary crack mask, first, the original images are divided into 448 × 448 overlapping image patches where these image patches are classified as cracks versus non-cracks using a deep CNN. Next, a modified U-Net is trained from scratch using only the crack patches for segmentation. A customized loss function that consists of binary cross entropy loss and the Dice loss is introduced to enhance the segmentation performance. Additionally, a Naïve Bayes fusion strategy is employed to integrate the crack score maps from different overlapping crack patches and to decide whether a pixel is crack or not. Comprehensive experiments have demonstrated that the proposed approach achieves an 81.71% mean intersection over union (mIoU) score across 5 different training/test splits, which is 7.29% higher than the baseline reference implemented with the original U-Net.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.10
• /
• pp.35-42
• /
• 2020

In this study, we propose a crack detection method using limited data with a U-Net based image inpainting technique that is a modified unsupervised anomaly detection method. Concrete cracking occurs due to a variety of causes and is a factor that can cause serious damage to the structure in the long term. In general, crack investigation uses an inspector's visual inspection on the concrete surfaces, which is less objective in judgment and has a high possibility of human error. Therefore, a method with objective and accurate image analysis processing is required. In recent years, the methods using deep learning have been studied to detect cracks quickly and accurately. However, when the amount of crack data on the building or infrastructure to be inspected is small, existing crack detection models using it often show a limited performance. Therefore, in this study, an unsupervised anomaly detection method was used to augment the data on the object to be inspected, and as a result of learning using the data, we confirmed the performance of 98.78% of accuracy and 82.67% of harmonic average (F1_Score).

• Journal of the Korean Society of Safety
• /
• v.34 no.6
• /
• pp.50-57
• /
• 2019

In this study, various types of deep learning models that have been proposed recently are classified according to data input / output types and analyzed to find the deep learning model suitable for constructing a crack detection model. First the deep learning models are classified into image classification model, object segmentation model, object detection model, and instance segmentation model. ResNet-101, DeepLab V2, Faster R-CNN, and Mask R-CNN were selected as representative deep learning model of each type. For the comparison, ResNet-101 was implemented for all the types of deep learning model as a backbone network which serves as a main feature extractor. The four types of deep learning models were trained with 500 crack images taken from real concrete structures and collected from the Internet. The four types of deep learning models showed high accuracy above 94% during the training. Comparative evaluation was conducted using 40 images taken from real concrete structures. The performance of each type of deep learning model was measured using precision and recall. In the experimental result, Mask R-CNN, an instance segmentation deep learning model showed the highest precision and recall on crack detection. Qualitative analysis also shows that Mask R-CNN could detect crack shapes most similarly to the real crack shapes.

• Journal of the Korean Geotechnical Society
• /
• v.34 no.12
• /
• pp.145-154
• /
• 2018

The damage investigation and inspection methods performed in concrete facilities such as bridges, tunnels, retaining walls and so on, are usually visually examined by the inspector using the surveying tool in the field. These methods highly depend on the subjectivity of the inspector, which may reduce the objectivity and reliability of the record. Therefore, the new image processing techniques are necessary in order to automatically detect the cracks and objectively analyze the characteristics of cracks. In this study, deep learning and image processing technique were developed to detect cracks and analyze characteristics in images for concrete facilities. Two-stage image processing pipeline was proposed to obtain crack segmentation and its characteristics. The performance of the method was tested using various crack images with a label and the results showed over 90% of accuracy on crack classification and segmentation. Finally, the crack characteristics (length and thickness) of the crack image pictured from the field were analyzed, and the performance of the developed technique was verified by comparing the actual measured values and errors.

• Journal of Korean Society of Steel Construction
• /
• v.22 no.6
• /
• pp.589-597
• /
• 2010

Steel structures have a higher probability of being damaged by fatigue than by other causes of deterioration. As such, their maintenance to prevent fatigue damage is essential to sustain their safety and performance during their service period. In their maintenance, the current state of their fatigue cracks must be assessed to determine appropriate reinforcement methods and the suitable time intervals of periodic inspections when fatigue cracks are detected. Determining the crack growth rate is a successful method of predicting fractures, but it requires technical knowledge on fracture mechanics and experience in numerical methods and software for finite element analysis. In this study, a fatigue crack growth test on through-thickness cracked steel plates was conducted to assess the crack growth rate without superior technical knowledge and experience. The relationship between the Crack Opening Displacement (COD) and the crack growth rate was found in relatively long fatigue cracks.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2009.05a
• /
• pp.441-442
• /
• 2009

Concrete has some self-healing ability if the crack is occurred. Concrete with a inorganic-organic chemical compound, self-healing ability is increased at the surface of the crack. In this study, we investigate self-healing performance of concrete using admixutre by performing permeability test.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2010.05a
• /
• pp.95-96
• /
• 2010

This paper presents the crack control of reinforcement concrete composite slabs which were produced with the extruded ECC panel. Cracking control performance was evaluated based on the flexural tests on real scale one-way slabs manufactured with or without ECC panel.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2002.05a
• /
• pp.23-28
• /
• 2002

This paper reports the performance results of hydration heat control of mass concrete walls with pipe cooling system. The thickness of walls ranged from 0.9 to 2.2m. In order to investigate the effect of pipe cooling on the thermal and thermal crack characteristics, the pipe cooling was conducted for 42 walls, and the investigation of thermal cracks was conducted for 14 walls. Based on the investigation, the pipe cooling method decreased the peak temperature of about 13-2$0^{\circ}C$ and the thermal crack width of about 30% for mass concrete walls.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.14 no.1
• /
• pp.112-118
• /
• 1990

Catastrophic fracture cannot be avoided after cracks(initiated from pre-existing defects) propagate rapidly with speeds comparable to a sound wave velocity of the materials. Preventing catastropic failure, crack arrest fracture toughness defined from dynamic(or kinetic) fracture mechanics point of view has been introduced in determining accurate and/or proper crack arrest fracture toughness of a material. For the past decades, many studies have been carried out to render proper theoretical and experimental backgrounds on the use of the static plain strain crack arrest fracture toughness, $K_{1a}$ (which seems to be a material property). $K_{1a}$ has been used to predict the performance of thick walled structures and has been considered as a measure of the ability of a material to stop a fast running crack. Determination of such a material property is of prime importance to the nuclear reactor pressure vessel and bridge materials industries. However, standards procedures for measuring toughness associated with fast running cracks are yet to exist. This study intends to give insight on the determination of the crack arrest fracture toughness of materials such as polymethylmethacrylate(PMMA), SM45C-steel, and A1 7075-T6. The effects of crack jump lengths and fast crack initiation stress intensity factor on the determination of $K_{1a}$ have been experimentally observed.erved.

• Steel and Composite Structures
• /
• v.26 no.1
• /
• pp.103-113
• /
• 2018

To study the effects of corner braces on fatigue performance of the U-rib and roof weld in steel bridge decks, the fatigue experiment was carried out to compare characteristics of the crack shape with and without corner braces. The improvement of fatigue life and stress variation after setting corner braces were also analysed. Different parameters of corner brace sizes, arrangements, and detail types were considered in the FEM models to obtain stress distribution and variation at the weld. Furthermore, enhancement of the fatigue performance by corner braces was evaluated. The results demonstrated that the corner brace could improve the fatigue life of the U-rib and roof weld, which exerted even no influence on the crack shape. Moreover, stress of the roof weld was decreased and the crack position was transferred from the root weld to U-rib and corner brace weld. It was suggested no weld scallop should be drilled on the corner brace. A transverse rib with lower height which was set between U-ribs was favourable for improvement of fatigue performance.