• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.33 no.4
• /
• pp.303-309
• /
• 2020

This study proposes a crack identification algorithm to analyze the surface condition of porcelain insulators and to efficiently visualize cracks. The proposed image processing algorithm for crack identification consists of two primary steps. In the first step, the brightness is eliminated by converting the image to the lab color space. Then, the background is removed by the K-means clustering method. After that, the optimum image treatment is applied using morphological image processing and median filtering to remove unnecessary noise, such as blobs. In the second step, the preprocessed image is converted to grayscale, and any cracks present in the image are identified. Next, the region properties, such as the number of pixels and the ratio of the major to the minor axis, are used to separate the cracks from the noise. Using this image processing algorithm, the precision of crack identification for all the sample images was approximately 80%, and the F1 score was approximately 70. Thus, this method can be helpful for efficient crack monitoring.

• Journal of the Korean Geotechnical Society
• /
• v.40 no.4
• /
• pp.137-144
• /
• 2024

The purpose of this paper is to analyze the significance of independent variables in estimating crack density using machine learning algorithms. The algorithms used were random forest and SHAP, with the independent variables being compressional wave velocity, shear wave velocity, porosity, and Poisson's ratio. Rock samples were collected from construction sites and processed into cylindrical forms to facilitate the acquisition of each input property. Artificial weathering was conducted twelve times to obtain values for both independent and dependent variables with multiple features. The application of the two algorithms revealed that porosity is a crucial independent variable in estimating crack density, whereas shear wave velocity has a relatively low impact. These results suggested that the four physical properties set as independent variables were sufficient for estimating crack density. Additionally, they presented a methodology for verifying the appropriateness of the independent variables using algorithms such as random forest and SHAP.

• Journal of Korean Association for Spatial Structures
• /
• v.22 no.2
• /
• pp.19-27
• /
• 2022

Visual inspection methods have limitations, such as reflecting the subjective opinions of workers. Moreover, additional equipment is required when inspecting the high-rise buildings because the height is limited during the inspection. Various methods have been studied to detect concrete cracks due to the disadvantage of existing visual inspection. In this study, a crack detection technology was proposed, and the technology was objectively and accurately through AI. In this study, an efficient method was proposed that automatically detects concrete cracks by using a Convolutional Neural Network(CNN) with the Orthomosaic image, modeled with the help of UAV. The concrete cracks were predicted by three different CNN models: AlexNet, ResNet50, and ResNeXt. The models were verified by accuracy, recall, and F1 Score. The ResNeXt model had the high performance among the three models. Also, this study confirmed the reliability of the model designed by applying it to the experiment.

• Journal of Korean Society of Water and Wastewater
• /
• v.27 no.6
• /
• pp.731-741
• /
• 2013

CCTV inspection method has been used in Korea for more than 20 years, but there is no proper assessment system for sewer failure severity that considers the domestic circumstances. This study classified the defects caused by the overburden load of reinforced concrete sewer pipes depending on severity and developed defect code by analyzing the domestic CCTV inspection videos. The defect score was assigned to each defect code, and it was classified into 5 grades for the decision-making of repair and rehabilitation. The result of this study is expected to be useful for domestic CCTV inspectors to assess the sewer condition and helpful for managers to make a decision of repair and rehabilitation.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.27 no.5
• /
• pp.113-119
• /
• 2023

Most infrastructure structures were completed during periods of economic growth. The number of infrastructure structures reaching their lifespan is increasing, and the proportion of old structures is gradually increasing. The functions and performance of these structures at the time of design may deteriorate and may even lead to safety accidents. To prevent this repercussion, accurate inspection and appropriate repair are requisite. To this end, demand is increasing for computer vision and deep learning technology to accurately detect even minute cracks. However, deep learning algorithms require a large number of training data. In particular, label images indicating the location of cracks in the image are required. To secure a large number of those label images, a lot of labor and time are consumed. To reduce these costs as well as increase detection accuracy, this study proposed a learning structure based on mean teacher method. This learning structure was trained on a dataset of 900 labeled image dataset and 3000 unlabeled image dataset. The crack detection network model was evaluated on over 300 labeled image dataset, and the detection accuracy recorded a mean intersection over union of 89.23% and an F1 score of 89.12%. Through this experiment, it was confirmed that detection performance was improved compared to supervised learning. It is expected that this proposed method will be used in the future to reduce the cost required to secure label images.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.21 no.4
• /
• pp.479-499
• /
• 2019

State assessment of an operational tunnel is usually done by performing visual inspection and durability tests by following the detailed guideline for safety inspection (SI) and/ or precision inspection for safety and diagnosis (PISD). In this study, 12 NATM tunnels, which have been operational for more than 10 years, were inspected to figure out the cause of longitudinal cracks for the purpose of modifying the scoring items in the state assessment NATM tunnel related to the longitudinal crack and the thickness of concrete lining. All investigated tunnels were classified into four groups depending on the shape and usage of each tunnel. The causes of longitudinal crack occurrence were analyzed by investigating the correlations between the longitudinal crack and the following four factors: the patterns of ground excavation; construction state of primary support system; characteristics of material properties of the concrete lining; and thickness of lining which was obtained by Ground Penetration Radar (GPR) tests. It was found that influencing factors causing longitudinal cracks in the lining were closely related with the construction condition of the primary support system, i.e. shotcrete, rockbolt, and steel-rib; crack occurrences were not much affected by the excavation patterns. As for the properties of concrete lining materials, occurrence of the longitudinal crack was mostly affected by the following three items: w/c ratio; contents of cement; and strength of lining. When estimating the lining thickness of the concrete lining by GPR tests and taking thickness effect into account in the statement assessment, it was concluded that increase of the index score by an average of 0.03 (ranging from 0.01 up to 0.071) is needed; a more realistic way of state assessment should be proposed in which the increased index score caused by lack of lining thickness should be taken into account.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.10
• /
• pp.1516-1522
• /
• 2017

We propose a novel monitoring system for diagnosing crack faults of the wind turbine using image information. The proposed method classifies a normal state and a abnormal state for the blade parts of the wind turbine. Specifically, the images are input to the proposed system in various states of wind turbine rotation. according to the blade condition. Then, the video of rotating blades on the wind turbine is divided into several image frames. Motion vectors are estimated using the previous and current images using the motion estimation, and the change of the motion vectors is analyzed according to the blade state. Finally, we determine the final blade state using the Support Vector Machine (SVM) classifier. In SVM, features are constructed using the area information of the blades and the motion vector values. The experimental results showed that the proposed method had high classification performance and its $F_1$ score was 0.9790.

• Journal of Korea Society of Industrial Information Systems
• /
• v.29 no.4
• /
• pp.1-11
• /
• 2024

Bridges crack and become damaged due to age and external factors such as earthquakes, lack of maintenance, and weather conditions. With the number of aging bridge on the rise, lack of maintenance can lead to a decrease in safety, resulting in structural defects and collapse. To prevent these problems and reduce maintenance costs, a system that can monitor the condition of bridge and respond quickly is needed. To this end, existing research has proposed artificial intelligence model that use sensor data to identify the location and extent of cracks. However, existing research does not use data from actual bridge to determine the performance of the model, but rather creates the shape of the bridge through simulation to acquire data and use it for training, which does not reflect the actual bridge environment. In this paper, we propose a bridge safety determination edge AI model that detects bridge abnormalities based on artificial intelligence by utilizing acceleration data from bridge occurring in the field. To this end, we newly defined filtering rules for extracting valid data from acceleration data and constructed a model to apply them. We also evaluated the performance of the proposed bridge safety determination edge AI model based on data collected in the field. The results showed that the F1-Score was up to 0.9565, confirming that it is possible to determine safety using data from real bridge, and that rules that generate similar data patterns to real impact data perform better.

• Hip & pelvis
• /
• v.36 no.2
• /
• pp.120-128
• /
• 2024

Purpose: The purpose of this study was to compare the clinical and radiographic outcomes with use of short-curved stems versus standard-length single wedged stems over a minimum follow-up period of five years. Materials and Methods: A retrospective study of primary total hip arthroplasties performed using the Fitmore^® stem (127 hips, 122 patients) and the M/L taper^® stem (195 hips, 187 patients) between October 2012 and June 2014 was conducted. The clinical and radiographic outcomes were obtained for evaluation over a minimum follow-up period of five years. Results: In both the Fitmore^® and M/L taper^® groups, the mean Harris hip score improved from 52.4 and 48.9 preoperatively to 93.3 and 94.5 at the final follow-up, respectively (P=0.980). The mean Western Ontario and McMaster Universities Osteoarthritis Index scores also improved from 73.3 and 76.8 preoperatively to 22.9 and 25.6 at the final follow-up, respectively (P=0.465). Fifteen hips (Fitmore^®: 14 hips; M/L taper^®: one hip, P<0.001) developed intraoperative cracks and were treated simultaneously with cerclage wiring. Radiography showed a radiolucent line in 24 hips in the Fitmore^® group and 12 hips in the M/L taper^® group (P=0.125). Cortical hypertrophy was detected in 29 hips (Fitmore^® group: 28 hips; M/L taper^® group: one hip, P<0.001). Conclusion: Similarly favorable clinical and radiographic outcomes were achieved with use of both short-curved stems and standard-length single wedged stems. However, higher cortical hypertrophy and a higher rate of femoral crack were observed with use of Fitmore^® stems.