• Journal of Korean Tunnelling and Underground Space Association
• /
• v.21 no.3
• /
• pp.419-432
• /
• 2019

Most of deep learning model training was proceeded by supervised learning, which is to train labeling data composed by inputs and corresponding outputs. Labeling data was directly generated manually, so labeling accuracy of data is relatively high. However, it requires heavy efforts in securing data because of cost and time. Additionally, the main goal of supervised learning is to improve detection performance for 'True Positive' data but not to reduce occurrence of 'False Positive' data. In this paper, the occurrence of unpredictable 'False Positive' appears by trained modes with labeling data and 'True Positive' data in monitoring of deep learning-based CCTV accident detection system, which is under operation at a tunnel monitoring center. Those types of 'False Positive' to 'fire' or 'person' objects were frequently taking place for lights of working vehicle, reflecting sunlight at tunnel entrance, long black feature which occurs to the part of lane or car, etc. To solve this problem, a deep learning model was developed by simultaneously training the 'False Positive' data generated in the field and the labeling data. As a result, in comparison with the model that was trained only by the existing labeling data, the re-inference performance with respect to the labeling data was improved. In addition, re-inference of the 'False Positive' data shows that the number of 'False Positive' for the persons were more reduced in case of training model including many 'False Positive' data. By training of the 'False Positive' data, the capability of field application of the deep learning model was improved automatically.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.5
• /
• pp.213-222
• /
• 2019

Determining the meaning of a keyword in a speech dialogue system is an important technology for the future implementation of an intelligent speech dialogue interface. After extracting keywords to grasp intention from user's utterance, the intention of utterance is determined by using the semantic mark of keyword. One keyword can have several semantic marks, and we regard the task of attaching the correct semantic mark to the user's intentions on these keyword as a problem of word sense disambiguation. In this study, about 23% of all keywords in the corpus is manually tagged to build a semantic mark dictionary, a synonym dictionary, and a context vector dictionary, and then the remaining 77% of all keywords is automatically tagged. The semantic mark of a keyword is determined by calculating the context vector similarity from the context vector dictionary. For an unregistered keyword, the semantic mark of the most similar keyword is attached using a synonym dictionary. We compare the performance of the system with manually constructed training set and semi-automatically expanded training set by selecting 3 high-frequency keywords and 3 low-frequency keywords in the corpus. In experiments, we obtained accuracy of 54.4% with manually constructed training set and 50.0% with semi-automatically expanded training set.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.33 no.3
• /
• pp.122-130
• /
• 2021

A wind-speed estimation at the arbitrary elevations is key component for the design of the offshore wind energy structures and the computation of the wind-wave generation. However, the wind-speed estimation of the target elevation has been carried out by using the typical functions and their typical parameters, e.g., power and logarithmic functions because the available wind speed data is limited to the specific elevation, such as 2~3m, 10 m, and so on. In this study, the parameters of the vertical profile functions are estimated with optimal and analyzed the parameter ranges using the HeMOSU-1 platform wind data monitored at the eight different locations. The results show that the mean value of the exponent of the power function is 0.1, which is significantly lower than the typically recommended value, 0.14. The values of the exponent, the friction velocity, and the roughness parameters are in the ranges 0.0~0.3, 0~10 (m/s), and 0.0~1.0 (m), respectively. The parameter ranges differ from the typical ranges because the atmospheric stability condition is assumed as the neutral condition. To improve the estimation accuracy, the atmospheric condition should be considered, and a more general (non-linear) vertical profile functions should be introduced to fit the diverse profile patterns and parameters.

• The Journal of the Korea Contents Association
• /
• v.21 no.1
• /
• pp.12-23
• /
• 2021

Due to high proportion of exports, Korea has a higher dependence on the global value chain (GVC) than other major developed countries. This reason, Korea has a structure that is sensitive to GVC changes. This is because Korean exports are concentrated on specific countries and items, and most of the materials for export tend to depend on imports. Currently, export restrictions resulting from trade disputes with Japan can affect the industry of Korea as a whole due to the supply of core materials. Therefore, in order to minimize economic damage caused by export regulations in the current situation, it is necessary to reorganize the GVC, through efforts to rapidly diversify imports and localize imports that depend on Japan. To this end, it is necessary to derive and classify imported goods that depend on Japan, and to localize items that are difficult to diversify imports, and prompt R&D investment is required for this. This study aims to support R&D investment policy through quantitative analysis based on big data rather than a decision-making method based on expert-centered qualitative analysis.

• The Journal of the Korea Contents Association
• /
• v.21 no.6
• /
• pp.35-43
• /
• 2021

Citrus quality is generally determined by its sugar content and acidity. In particular, sugar content is a very important factor because it determines the taste of citrus. Currently, the most commonly used method of measuring citrus sugar content in farms is a portable juiced sugar meter and a non-destructive sugar meter. This method can be easily measured by individuals, but the accuracy of the sugar content is inferior to that of the citrus NongHyup official machine. In particular, there is an error difference of 0.5 Brix or more, which is still insufficient for use in the field. Therefore, in this paper, we propose an AI model that predicts the citrus sugar content of unmeasured days within the error range of 0.5 Brix or less based on the previously collected citrus sugar content and meteorological data (average temperature, humidity, rainfall, solar radiation, and average wind speed). In addition, it was confirmed that the prediction model proposed through performance evaluation had an mean absolute error of 0.1154 for Seongsan area and 0.1983 for the Hawon area in Jeju Island. Lastly, the proposed model supports an error difference of less than 0.5 Brix and is a technology that supports predictive measurement, so it is expected that its usability will be highly progressive.

• Resources Recycling
• /
• v.30 no.3
• /
• pp.70-81
• /
• 2021

This study evaluated the recyclability of waste plastics in used coffee capsules disposed of as municipal waste. For recycling, a new material separation process was developed to remove the coffee grounds through primary crushing, washing, sieving, and secondary crushing, followed by corona discharge electrostatic separation. Furthermore, for the under 10 mm size fraction samples, the aluminum removal and the plastic recovery were 95.4% and 98.3%, respectively, under optimal conditions. In addition, for the 15 mm fraction samples, the aluminum removal and the plastic recovery were 91.3% and 97.2%, respectively. To evaluate the recyclability of the separated waste plastics, the samples were pelleted, and their material properties were analyzed. No hazardous substances were detected, and the results were similar to those for homo-PP. Therefore, it was confirmed tha t sufficient functiona lity existed a s recycled PP. However, owing to the da rk color of the pellets, limited applications to black or dark products are expected.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.33 no.6
• /
• pp.217-225
• /
• 2021

In this study, wave spectrum data were calculated using the water surface elevation data observed at 5Hz intervals from the HeMOSU-2 meteorological tower installed on the west coast of Korea, and wave parameters were estimated using wave spectrum data. For all significant wave height ranges, the peak enhancement parameter (γ_opt) of the JONSWAP spectrum and the scale parameter (α) and shape parameter (β) of the modify BM spectrum were estimated based on the observed spectrum, and the distribution of each parameter was confirmed. As a result of the analysis, the peak enhancement parameter (γ_opt) of the JONSWAP spectrum was calculated to be 1.27, which is very low compared to the previously proposed 3.3. And in the range of all significant wave heights, the distribution of the peak enhancement parameter (γ_opt) was shown as a combined distribution of probability mass function (PMF) and probability density function (PDF). In addition, the scale parameter (α) and shape parameter (β) of the modify BM spectrum were estimated to be [0.245, -1.278], which are lower than the existing [0.300, -1.098], and the result of the linear correlation analysis between the two parameters was β = -3.86α.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.35 no.2
• /
• pp.109-117
• /
• 2022

There has been an increasing number of studies on photovoltaic energy generation system in an offshore site with the largest energy generation efficiency, as increasing the researches and developments of renewable energies for use of offshore space and resources to replace existing fossil fuels and resolve environmental challenges. For installation and operation of floating photovoltaic systems in an offshore site with harsher environmental conditions, a stiffness of structural members comprising the total system must be reinforced to inland water spaces as dams, reservoirs etc., which have relatively weak condition. However, there are various limitations for the reinforcement of structural stiffness of the system, including producible size, total mass of the system, economic efficiency, etc. Thus, in this study, a floating breakwater is considered for reducing wave loads on the system and minimizing the reinforcement of the structural members. Wave reduction performances of floating breakwaters are evaluated, considering size and distance to the system. The wave loads on the system are evaluated using the higher-order boundary element method (HOBEM), considering the multi-body effect of buoys. Stresses on structural members are assessed by coupled analyses using the finite element method (FEM), considering the wave loads and hydrodynamic characteristics. As the maximum stresses on each of the cases are reviewed and compared, the effect of floating breakwater for floating photovoltaic system is checked, and it is confirmed that the size of breakwater has a significant effect on structural responses of the system.

• The Journal of the Korea Contents Association
• /
• v.22 no.10
• /
• pp.11-21
• /
• 2022

All-solid-state batteries are one of the promising candidates for next-generation batteries and are drawing attention as a key component that will lead the future electric vehicle industry. This study analyzes 10,280 comments on Reddit, which is a global social media, in order to identify policy issues and public interest related to all-solid-state batteries from 2016 to 2021. Text mining such as frequency analysis, association rule analysis, and topic modeling, and sentiment analysis are applied to the collected global data to grasp global trends, compare them with the South Korean government's all-solid-state battery development strategy, and suggest policy directions for its national research and development. As a result, the overall sentiment toward all-solid-state battery issues was positive with 50.5% positive and 39.5% negative comments. In addition, as a result of analyzing detailed emotions, it was found that the public had trust and expectation for all-solid-state batteries. However, feelings of concern about unresolved problems coexisted. This study has an academic and practical contribution in that it presented a text mining analysis method for deriving key issues related to all-solid-state batteries, and a more comprehensive trend analysis by employing both a top-down approach based on government policy analysis and a bottom-up approach that analyzes public perception.

• Ecology and Resilient Infrastructure
• /
• v.8 no.4
• /
• pp.212-222
• /
• 2021

In this study, the flow resistance coefficient was calculated considering the physical properties and distribution characteristics of floodplain vegetation, and the effect of floodplain vegetation distribution on flow characteristics was analyzed by reflecting it in a two-dimensional numerical simulation. The three-dimensional point clouds of vegetation acquired using ground lidar were analyzed to apply floodplain vegetation's physical properties to the existing formula for vegetation flow resistance calculation. The floodplain vegetation distribution in the modeling was divided into locally distributed and fully distributed conditions in the floodplain. As a result of the simulation of the study site, the flow resistance coefficient of floodplain vegetation was found to have a value of about five times or more compared to the flow resistance coefficient of the main channel bed when the design flood occurs based on Manning's n coefficient. Also, it affected the hydraulic characteristics in the main channel and floodplain.