• Journal of the Korea Academia-Industrial cooperation Society
• /
• 제18권4호
• /
• pp.487-493
• /
• 2017

The main purpose of this study is to evaluate the main parameters involved in the asphalt pavement distresses, including IRI (International Rough Index), fatigue, and permanent deformation. The main parameters are the region (Seoul and Busan), traffic level, asphalt binder, maximum aggregate of surface course, thickness of the surface course and base. A total of 64 case studies were carried out under the auspices of the KPRP (Korea Pavement Research Program). From the analysis of the KPRP test results, the key factors for the asphalt pavement distress were determined. Considering the effect of one variable in the basic condition, asphalt binder was the major factor having an effect on the distresses for an AADT (Annual Average Daily Traffic) of 5000 in the Seoul area. Among the remaining factors, the results were found to be in the order of the base layer thickness (A), surface layer thickness (B), and aggregate particle size thickness (D). The same results were obtained for an AADT of 10000. In the case of Busan with an AADT of 5000, the same result was obtained as for Seoul. Among the remaining factors, the results were in the order of the base layer thickness (A), aggregate particle thickness (D), and surface layer thickness (B). Even though there was a slight difference in the effect of the traffic level and region, asphalt binder was the parameter having the greatest effect on the asphalt pavement distress. In the case where the effect of multiple parameters was analyzed, the combination of the asphalt binder and base thickness showed a relatively strong effect.

• Journal of Korean Tunnelling and Underground Space Association
• /
• 제3권3호
• /
• pp.15-29
• /
• 2001

For the pertinent use of NMT method, both characteristics of joints (JRC, JCS and ${\phi}_r$) and characteristics of rock mass (Q-Value) must be investigated carefully. The main objective of the study presented is to investigate how sensitive the predicted behaviour of an underground excavation is to various realistic assumptions about some input parameter for the jointed rock mass. Joint pattern in the tunnel is predicted by statistical approach (chi-square test). In this paper, sensitivity studies involving in joint characteristics were carried out. The parametric studies involving change in Barton-Bandis joint model have shown that JCS is relatively insensitive to JRC and ${\phi}_r$. An increase in JRC value may not, according to the Barton-Bandis model, necessarily lead to a decrease in displacement. The importance of dilation in predicting the behaviour of a rock mass around an excavation is emphasized from a comparison of the Barton-Bandis joint behaviour model with the Mohr-Coulomb model. The Barton-Bandis model predicted higher stress, which allow for the build-up of stress caused by dilatant behaviour.

• Proceedings of the KSR Conference
• /
• 한국철도학회 2008년도 추계학술대회 논문집
• /
• pp.1069-1077
• /
• 2008

The railway abrasion measurement system have to satisfy two conditions to increase the measurement accuracy as follows. The laser region which is projected on the rail have to be extracted without the geometrical distortion. The mapping of the acquired laser region data on the rail profile have to be processed exactly. But, the conventional railway abrasion measurement system is deeply effected by the foreign substance( dust, rainwater, and so on ) on the railway or the sensitive response characteristic of the laser to the external measurement circumstance, and then the measurement errors arise from above factors. When the laser region is projected on the rail extracts from the acquired image, the interference of the light with the same frequency as the laser system occurs the serious problems. In the process of the mapping between the railway profile and the extracted laser region, the measurement accuracy is very highly effected by the geometrical distortion and the abnormal variation. In this Paper, we propose the novel method to increase the accuracy of the railway abrasion measurement dramatically. we designed and manufactured the high precision and fast image processing board with DSP Core and FPGA to measure the railway abrasion. The image processing board has the capability that the image of 1024X1280 from camera can be processed with the speed of 480 frame/sec. And, we apply the image processing algorithm base on the wavelet to extract the laser region is projected on the rail exactly. Finally, we developed high precision railway abrasion measurement system with the error range less than +/-0.5mm by which 2D image data is covered 3D data and mapped on the rail profile using the camera model and the perspective transform.

• Journal of Convergence for Information Technology
• /
• 제11권11호
• /
• pp.57-65
• /
• 2021

We propose a novel method to detect abnormal data of specific symptoms using deep learning in air pollution measurement system. Existing methods generally detect abnomal data by classifying data showing unusual patterns different from the existing time series data. However, these approaches have limitations in detecting specific symptoms. In this paper, we use DeepLab V3+ model mainly used for foreground segmentation of images, whose structure has been changed to handle one-dimensional data. Instead of images, the model receives time-series data from multiple sensors and can detect data showing specific symptoms. In addition, we improve model's performance by reducing the complexity of noisy form time series data by using 'piecewise aggregation approximation'. Through the experimental results, it can be confirmed that anomaly data detection can be performed successfully.

• Journal of Bio-Environment Control
• /
• 제28권3호
• /
• pp.204-211
• /
• 2019

The purpose of this study is to provide basic data for setting environmental design standards for domestic greenhouses. We conducted experiments on thermal environment measurement at two commercial greenhouses where hot water heating system is adopted. We analyzed heat transfer characteristics of hot water heating pipes and heat emission per unit length of heating pipes was presented. The average air temperature in two greenhouses was controlled to $16.3^{\circ}C$ and $14.6^{\circ}C$ during the experiment, respectively. The average water temperature in heating pipes was $52.3^{\circ}C$ and $45.0^{\circ}C$, respectively. Experimental results showed that natural convection heat transfer coefficient of heating pipe surface was in the range of $5.71{\sim}7.49W/m^2^{\circ}C$. When the flow rate in heating pipe was 0.5m/s or more, temperature difference between hot water and pipe surface was not large. Based on this, overall heat transfer coefficient of heating pipe was derived as form of laminar natural convection heat transfer coefficient in the horizontal cylinder. By modifying the equation of overall heat transfer coefficient, a formula for calculating the heat emission per unit length of hot water heating pipe was developed, which uses pipe size and temperature difference between hot water and indoor air as input variables. The results of this study were compared with domestic and foreign data, and it was found to be closest to JGHA data. The data of NAAS, BALLS and ASHRAE were judged to be too large. Therefore, in order to set up environmental design standards for domestic greenhouses, it is necessary to fully examine those data through further experiments.

• Database Research
• /
• 제34권3호
• /
• pp.148-158
• /
• 2018

The optical character recognition (OCR) is a technique to extract and recognize texts from images. It is an important preprocessing step in data analysis since most actual text information is embedded in images. Many OCR engines have high recognition accuracy for images where texts are clearly separable from background, such as white background and black lettering. However, they have low recognition accuracy for images where texts are not easily separable from complex background. To improve this low accuracy problem with complex images, it is necessary to transform the input image to make texts more noticeable. In this paper, we propose a method to segment an input image into text lines to enable OCR engines to recognize each line more efficiently, and to determine the final output by comparing the recognition rates of CLAHE module and Two-step module which distinguish texts from background regions based on image processing techniques. Through thorough experiments comparing with well-known OCR engines, Tesseract and Abbyy, we show that our proposed method have the best recognition accuracy with complex background images.

• Journal of Internet Computing and Services
• /
• 제22권2호
• /
• pp.1-9
• /
• 2021

Traditionally, most malicious codes have been analyzed using feature information extracted by domain experts. However, this feature-based analysis method depends on the analyst's capabilities and has limitations in detecting variant malicious codes that have modified existing malicious codes. In this study, we propose a ResNet-Variational AutoEncder-based variant malware classification method that can classify a family of variant malware without domain expert intervention. The Variational AutoEncoder network has the characteristics of creating new data within a normal distribution and understanding the characteristics of the data well in the learning process of training data provided as input values. In this study, important features of malicious code could be extracted by extracting latent variables in the learning process of Variational AutoEncoder. In addition, transfer learning was performed to better learn the characteristics of the training data and increase the efficiency of learning. The learning parameters of the ResNet-152 model pre-trained with the ImageNet Dataset were transferred to the learning parameters of the Encoder Network. The ResNet-Variational AutoEncoder that performed transfer learning showed higher performance than the existing Variational AutoEncoder and provided learning efficiency. Meanwhile, an ensemble model, Stacking Classifier, was used as a method for classifying variant malicious codes. As a result of learning the Stacking Classifier based on the characteristic data of the variant malware extracted by the Encoder Network of the ResNet-VAE model, an accuracy of 98.66% and an F1-Score of 98.68 were obtained.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• 제25권10호
• /
• pp.1409-1415
• /
• 2021

The Gysel divider has the advantage of easily setting the resistor in the circuit. If the line impedance in the Gysel divider is set differently, the input signal can be distributed to the two output ports at various distribution ratios. This paper proposes the Gysel divider that can change the power distribution to 1:3 or 3:1 by changing the line impedance. The impedance change of the line can be implemented by placing a floating copper plate on the bottom of the microstrip-line. When the floating copper plate and the ground plane are connected, the line operates as the microstrip-line, and when the floating copper plate and the ground plane are disconnected, the line operates as the coplanar-line. The proposed Gysel divider was fabricated at the center frequency of 1.5GHz. The fabricated 3:1 Gysel divider has a stable value S₁₁ of below -17dB, S₂₁/S₃₁ of 4.8±0.2dB, S₂₁(to high output port) of -1.39±0.12dB and S₃₁(to low output port) of -6.15±0.08dB over 1.3~1.7GHz.

• Journal of the Korea Computer Graphics Society
• /
• 제28권2호
• /
• pp.1-9
• /
• 2022

In this paper, we propose a vector-based augmentation technique that can generate data required for crack detection and a ConvNet(Convolutional Neural Network) technique that can learn it. Detecting cracks quickly and accurately is an important technology to prevent building collapse and fall accidents in advance. In order to solve this problem with artificial intelligence, it is essential to obtain a large amount of data, but it is difficult to obtain a large amount of crack data because the situation for obtaining an actual crack image is mostly dangerous. This problem of database construction can be alleviated with elastic distortion, which increases the amount of data by applying deformation to a specific artificial part. In this paper, the improved crack pattern results are modeled using ConvNet. Rather than elastic distortion, our method can obtain results similar to the actual crack pattern. By designing the crack data augmentation based on a vector, rather than the pixel unit used in general data augmentation, excellent results can be obtained in terms of the amount of crack change. As a result, in this paper, even though a small number of crack data were used as input, a crack database can be efficiently constructed by generating various crack directions and patterns.

• Journal of Internet Computing and Services
• /
• 제24권4호
• /
• pp.25-36
• /
• 2023

Today, as AI (Artificial Intelligence) technology is introduced in various fields, including security, the development of technology is accelerating. However, with the development of AI technology, attack techniques that cleverly bypass malicious behavior detection are also developing. In the classification process of AI models, an Adversarial attack has emerged that induces misclassification and a decrease in reliability through fine adjustment of input values. The attacks that will appear in the future are not new attacks created by an attacker but rather a method of avoiding the detection system by slightly modifying existing attacks, such as Adversarial attacks. Developing a robust model that can respond to these malware variants is necessary. In this paper, we propose two methods of generating Adversarial attacks as efficient Adversarial attack generation techniques for improving Robustness in AI models. The proposed technique is the XAI-based attack technique using the XAI technique and the Reference based attack through the model's decision boundary search. After that, a classification model was constructed through a malicious code dataset to compare performance with the PGD attack, one of the existing Adversarial attacks. In terms of generation speed, XAI-based attack, and reference-based attack take 0.35 seconds and 0.47 seconds, respectively, compared to the existing PGD attack, which takes 20 minutes, showing a very high speed, especially in the case of reference-based attack, 97.7%, which is higher than the existing PGD attack's generation rate of 75.5%. Therefore, the proposed technique enables more efficient Adversarial attacks and is expected to contribute to research to build a robust AI model in the future.