• Structural Engineering and Mechanics
• /
• v.91 no.5
• /
• pp.443-457
• /
• 2024

At present, the traditional concrete surface inspection methods based on artificial vision have the problems of high cost and insecurity, while the computer vision methods rely on artificial selection features in the case of sensitive environmental changes and difficult promotion. In order to solve these problems, this paper introduces deep learning technology in the field of computer vision to achieve automatic feature extraction of structural damage, with excellent detection speed and strong generalization ability. The main contents of this study are as follows: (1) A method based on DeepLabV3+ convolutional neural network model is proposed for surface detection of post-earthquake structural damage, including surface damage such as concrete cracks, spaling and exposed steel bars. The key semantic information is extracted by different backbone networks, and the data sets containing various surface damage are trained, tested and evaluated. The intersection ratios of 54.4%, 44.2%, and 89.9% in the test set demonstrate the network's capability to accurately identify different types of structural surface damages in pixel-level segmentation, highlighting its effectiveness in varied testing scenarios. (2) A semantic segmentation model based on DeepLabV3+ convolutional neural network is proposed for the detection and evaluation of post-earthquake structural components. Using a dataset that includes building structural components and their damage degrees for training, testing, and evaluation, semantic segmentation detection accuracies were recorded at 98.5% and 56.9%. To provide a comprehensive assessment that considers both false positives and false negatives, the Mean Intersection over Union (Mean IoU) was employed as the primary evaluation metric. This choice ensures that the network's performance in detecting and evaluating pixel-level damage in post-earthquake structural components is evaluated uniformly across all experiments. By incorporating deep learning technology, this study not only offers an innovative solution for accurately identifying post-earthquake damage in civil engineering structures but also contributes significantly to empirical research in automated detection and evaluation within the field of structural health monitoring.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.37 no.9
• /
• pp.1175-1182
• /
• 2013

The lock-in mode infrared thermography (IRT) technique has been developed to improve the detection capability of defects in materials with high thermal conductivity, and it has been shown to provide better detection capability than conventional active IRT. Therefore, to investigate the application of this technique to nuclear piping components, lock-in mode IRT tests were conducted on pipe specimens containing simulated wall-thinning defects. Phase images of the wall-thinning defects were obtained from the tests, and they were compared with thermal images obtained from conventional active IRT tests. It showed that the ability to size the detected wall-thinning defects in piping components was improved by using lock-in mode IRT. The improvement was especially apparent when detecting short and narrow defects and defects with slanted edges. However, the detection capability for shallow wall-thinning defects did not improve much when using lock-in mode IRT.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.8
• /
• pp.3865-3883
• /
• 2016

Infrastructure as a Service (IaaS) encapsulates computer hardware into a large amount of virtual and manageable instances mainly in the form of virtual machine (VM), and provides rental service for users. Currently, VM anomaly incidents occasionally occur, which leads to performance issues and even downtime. This paper aims at detecting anomalous VMs based on performance metrics data of VMs. Due to the dynamic nature and increasing scale of IaaS, detecting anomalous VMs from voluminous correlated and non-Gaussian monitored performance data is a challenging task. This paper designs an anomaly detection framework to solve this challenge. First, it collects 53 performance metrics to reflect the running state of each VM. The collected performance metrics are testified not to follow the Gaussian distribution. Then, it employs independent components analysis (ICA) instead of principal component analysis (PCA) to extract independent components from collected non-Gaussian performance metric data. For anomaly detection, it employs multi-class Bayesian classification to determine the current state of each VM. To evaluate the performance of the designed detection framework, four types of anomalies are separately or jointly injected into randomly selected VMs in a campus-wide testbed. The experimental results show that ICA-based detection mechanism outperforms PCA-based and LDA-based detection mechanisms in terms of sensitivity and specificity.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2003.10a
• /
• pp.379-384
• /
• 2003

In this paper, we introduce a change detection system that can extract and analyze change elements from high-resolution satellite imagery as well as low- or middle-resolution satellite imagery. The developed system provides not only 7 pixel-based methods that can be used to detect change from low- or middle-resolution satellite images but also a float window concept that can be used in manual change detection from high-resolution satellite images. This system enables fast process of the very large image, because it is constituted by OGC grid coverage components. Also new change detection algorithms can be easily added into this system if once they are made into grid coverage components.

• Journal of Information Processing Systems
• /
• v.9 no.1
• /
• pp.89-102
• /
• 2013

The paper presents a fuzzy based impulse noise filter for both gray scale and color images. The proposed approach is based on the technique of boundary discriminative noise detection. The algorithm is a multi-step process comprising detection, filtering and color correction stages. The detection procedure classifies the pixels as corrupted and uncorrupted by computing decision boundaries, which are fuzzified to improve the outputs obtained. In the case of color images, a correction term is added by examining the interactions between the color components for further improvement. Quantitative and qualitative analysis, performed on standard gray scale and color image, shows improved performance of the proposed technique over existing state-of-the-art algorithms in terms of Peak Signal to Noise Ratio (PSNR) and color difference metrics. The analysis proves the applicability of the proposed algorithm to random valued impulse noise.

• International Journal of Computer Science & Network Security
• /
• v.22 no.5
• /
• pp.97-102
• /
• 2022

Network intrusion detection is becoming an increasing necessity for both organizations and individuals alike. Detecting intrusions is one of the major components that aims to prevent information compromise. Automated systems have been put to use due to the voluminous nature of the domain. The major challenge for automated models is the noise and data imbalance components contained in the network transactions. This work proposes an ensemble model, Attribute Subset Selector Bagging (ASUB) that can be used to effectively handle noise and data imbalance. The proposed model performs attribute subset based bag creation, leading to reduction of the influence of the noise factor. The constructed bagging model is heterogeneous in nature, hence leading to effective imbalance handling. Experiments were conducted on the standard intrusion detection datasets KDD CUP 99, Koyoto 2006 and NSL KDD. Results show effective performances, showing the high performance of the model.

• ETRI Journal
• /
• v.38 no.6
• /
• pp.1074-1084
• /
• 2016

This paper presents a transmission method for improving human body communications in terms of spectral efficiency, and the performances of bit-error-rate (BER) and frame synchronization, with a highly simplified structure. Compared to the conventional frequency selective digital transmission supporting IEEE standard 802.15.6 for wireless body area networks, the proposed scheme improves the spectral efficiency from 0.25 bps/Hz to 1 bps/Hz based on the 3-dB bandwidth of the transmit spectral mask, and the signal-to-noise-ratio (SNR) by 0.51 dB at a BER of $10^{-6}$ with an 87.5% reduction in the detection complexity of the length of the Hamming distance computation. The proposed preamble structure using its customized detection algorithm achieves perfect frame synchronization at the SNR of a BER of $10^{-6}$ by applying the proposed pre-processing to compensate for the distortions on the preamble signals due to the band-limit effects by transmit and receive filters.

• Journal of Korea Multimedia Society
• /
• v.9 no.1
• /
• pp.11-22
• /
• 2006

Recently face detection has actively been researched due to its wide range of applications, such as personal identification and security systems. In this paper, a new face detection method based on the distribution map is proposed. Face-like regions are first extracted by applying the skin color map with the frequency to a color image and then, possible eye regions are determined by using the pupil color distribution map within the face-like regions. This enables the reduction of space for finding facial features. Eye candidates are detected by means of a template matching method using weighted window, which utilizes the correlation values of the luminance component and chrominance components as feature vectors. Finally, a cost function for mouth detection and location information between the facial features are applied to each pair of the eye candidates for face detection. Experimental results show that the proposed method can achieve a high performance.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.51 no.10
• /
• pp.495-501
• /
• 2002

A difference based current transformer (CT) saturation detection algorithm uses the third difference of a secondary current to detect the instants of the beginning/end of saturation. The third difference of a secondary current contains high frequency components when a CT is saturated. Thus, an effect of an anti-aliasing lowpass filter implemented in digital protection relays on the detection algorithm should be studied. This paper describes performance analysis of a lowpass filter on the CT saturation detection algorithm. The cutoff frequency of the lowpass filter is normally set to be half of a sampling frequency. In this Paper, two sampling frequencies of 3,840 (Hz) corresponding to 64 sample/cycle (s/c) and 1,920 (Hz) corresponding to 32 (s/c) are studied; the cutoff frequencies of the lowpass filters are set to be 1,920 (Hz), 960 (Hz) and 960(Hz), 480(Hz), respectively. And the proposed algorithm is verified by experiment. A 2nd order Butterworth filter is designed as a lowpass filter. The test results and experiment results clearly indicate that the saturation detection algorithm successfully detects the instants of the beginning/end of saturation even though a secondary current is filtered by the designed lowpass filters.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.43 no.2
• /
• pp.33-38
• /
• 2020

Anomaly detection of Machine Learning such as PCA anomaly detection and CNN image classification has been focused on cross-sectional data. In this paper, two approaches has been suggested to apply ML techniques for identifying the failure time of big time series data. PCA anomaly detection to identify time rows as normal or abnormal was suggested by converting subjects identification problem to time domain. CNN image classification was suggested to identify the failure time by re-structuring of time series data, which computed the correlation matrix of one minute data and converted to tiff image format. Also, LASSO, one of feature selection methods, was applied to select the most affecting variables which could identify the failure status. For the empirical study, time series data was collected in seconds from a power generator of 214 components for 25 minutes including 20 minutes before the failure time. The failure time was predicted and detected 9 minutes 17 seconds before the failure time by PCA anomaly detection, but was not detected by the combination of LASSO and PCA because the target variable was binary variable which was assigned on the base of the failure time. CNN image classification with the train data of 10 normal status image and 5 failure status images detected just one minute before.