• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.261-261
• /
• 2020

물관리의 기본이 되는 연속적인 유량 자료 확보를 위해서는 정확도 높은 수위-유량 관계 곡선식 개발이 필수적이다. 수위-유량 관계곡선식은 모든 수문시설 설계의 기초가 되며 홍수, 가뭄 등 물재해 대응을 위해서도 중요한 의미를 가지고 있다. 그러나 일반적으로 유량 측정은 많은 비용과 시간이 들고, 식생성장, 단면변화 등의 통제특성(control)이 변함에 따라 구간분리, 기간분리와 같은 비선형적인 양상이 나타나 자료 해석에 어려움이 존재한다. 특히, 국내 하천의 경우 자연적 및 인위적인 환경 변화가 다양하여 지점 및 기간에 따라 세밀한 분석이 요구된다. 머신러닝(Machine Learning)이란 데이터를 통해 컴퓨터가 스스로 학습하여 모델을 구축하고 성능을 향상시키는 일련의 과정을 뜻한다. 기존의 수위-유량 관계곡선식은 개발자의 판단에 의해 데이터의 종류와 기간 등을 설정하여 회귀식의 파라미터를 산출한다면, 머신러닝은 유효한 전체 데이터를 이용해 스스로 학습하여 자료 간 상관성을 찾아내 모델을 구축하고 성능을 지속적으로 향상 시킬 수 있다. 머신러닝은 충분한 수문자료가 확보되었다는 전제 하에 복잡하고 가변적인 수자원 환경을 반영하여 유량 추정의 정확도를 지속적으로 향상시킬 수 있다는 이점을 가지고 있다. 본 연구는 머신러닝의 대표적인 알고리즘들을 활용하여 유량을 추정하는 모델을 구축하고 성능을 비교·분석하였다. 대상지역은 안정적인 수량을 확보하고 있는 한강수계의 거운교 지점이며, 사용자료는 2010~2018년의 시간, 수위, 유량, 수면폭 등 이다. 프로그램은 파이썬을 기반으로 한 머신러닝 라이브러리인 사이킷런(sklearn)을 사용하였고 알고리즘은 랜덤포레스트 회귀, 의사결정트리, KNN(K-Nearest Neighbor), rgboost을 적용하였다. 학습(train) 데이터는 입력자료 종류별로 조합하여 6개의 세트로 구분하여 모델을 구축하였고, 이를 적용해 검증(test) 데이터를 RMSE(Roog Mean Square Error)로 평가하였다. 그 결과 모델 및 입력 자료의 조합에 따라 3.67~171.46로 다소 넓은 범위의 값이 도출되었다. 그 중 가장 우수한 유형은 수위, 연도, 수면폭 3개의 입력자료를 조합하여 랜덤포레스트 회귀 모델에 적용한 경우이다. 비교를 위해 동일한 검증 데이터를 한국수문조사연보(2018년) 내거운교 지점의 수위별 수위-유량 곡선식을 이용해 유량을 추정한 결과 RMSE가 3.76이 산출되어, 머신러닝이 세분화된 수위-유량 곡선식과 비슷한 수준까지 성능을 내는 것으로 확인되었다. 본 연구는 양질의 유량자료 생산을 위해 기 구축된 수문자료를 기반으로 머신러닝 기법의 적용 가능성을 검토한 기초 연구로써, 국내 효율적인 수문자료 측정 및 수위-유량 곡선 산출에 도움이 될 수 있을 것으로 판단된다. 향후 수자원 환경 및 통제특성에 영향을 미치는 다양한 영향변수를 파악하기 위해 기상자료, 취수량 등의 입력 자료를 적용할 필요가 있으며, 머신러닝 내 비지도학습인 딥러닝과 같은 보다 정교한 모델에 대한 추가적인 연구도 수행되어야 할 것이다.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.28 no.6B
• /
• pp.665-673
• /
• 2008

This study is to assess the future potential climate and land use change impact on streamflow and stream water quality of the study watershed using the established model parameters (I). The CCCma (Canadian Centre for Climate Modelling and Analysis) CGCM2 (Canadian Global Coupled Model) based on IPCC SRES (Special Report Emission Scenarios) A2 and B2 scenarios were adopted for future climate condition, and the data were downscaled by Stochastic Spatio-Temporal Random Cascade Model technique. The future land use condition was predicted by using modified CA-Markov (Cellular Automata-Markov chain) technique with the past time series of Landsat satellite images. The model was applied for the future extreme precipitation cases of around 2030, 2060 and 2090. The predicted results showed that the runoff ratio increased 8% based on the 2005 precipitation (1160.1 mm) and runoff ratio (65%). Accordingly the Sediment, T-N and T-P also increased 120%, 16% and 10% respectively for the case of 50% precipitation increase. This research has the meaning in providing the methodological procedures for the evaluation of future potential climate and land use changes on watershed hydrology and stream water quality. This model result are expected to plan in advance for healthy and sustainable watershed management and countermeasures of climate change.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.25 no.4
• /
• pp.311-325
• /
• 2023

This study was conducted to analyze the habitat suitability of the major honey trees including Kalopanax septemlobus Koidz., Prunus spp., Tilia spp., and Styrax obassia Siebold & Zucc. indigenous to mountain Gariwang and Yumeong using the machine learning approach (i.e., MaxEnt model). The AUC values of the model predictions were mostly above 0.7, and the results of the response curves showed that the environmental drivers that had effects on the habitat suitability of the major honey trees were elevation, mean annual precipitation, and mean annual temperature. These results indicate that climatic drivers along the elevation gradient are the main environmental drivers in explaining the distribution patterns of the major honey trees. In addition, the results of the response curves of Prunus spp. and Styrax obassia Siebold & Zucc. differed slightly in terms of slope and mean annual solar radiation as the main environmental drivers. The results of this study will be valuable for the establishment of honey tree forests and management plans for the natural and artificial forests in South Korea, as well as for the mapping the distribution of honey trees. Further studies at different regional levels, reflecting biotic drivers, will be needed to expand the production of honey and pollen at different strata and to produce honey annually.

• Tunnel and Underground Space
• /
• v.34 no.3
• /
• pp.208-217
• /
• 2024

There is so much research effort for developing and implementing deep learning-based surveillance systems to manage health and safety issues in construction sites. Especially, the development of deep learning-based object detection in various environmental changes has been progressing because those affect decreasing searching performance of the model. Among the various environmental variables, the accuracy of the object detection model is significantly dropped under low illuminance, and consistent object detection accuracy cannot be secured even the model is trained using low-light images. Accordingly, there is a need of low-light enhancement to keep the performance under low illuminance. Therefore, this paper conducts a comparative study of various deep learning-based low-light image enhancement models (GLADNet, KinD, LLFlow, Zero-DCE) using the acquired construction site image data. The low-light enhanced image was visually verified, and it was quantitatively analyzed by adopting image quality evaluation metrics such as PSNR, SSIM, Delta-E. As a result of the experiment, the low-light image enhancement performance of GLADNet showed excellent results in quantitative and qualitative evaluation, and it was analyzed to be suitable as a low-light image enhancement model. If the low-light image enhancement technique is applied as an image preprocessing to the deep learning-based object detection model in the future, it is expected to secure consistent object detection performance in a low-light environment.

• Journal of Nutrition and Health
• /
• v.42 no.1
• /
• pp.78-96
• /
• 2009

The purpose of this study was to analyze and evaluate the menu served in government offices foodservice by using Kasavana & Smith's Menu-Engineering. Sales and food costs were collected from the daily sales reports for a year from Jan 2 to Dec 31 in 2007. Calculation for menu analysis and customer's data were done by computer using the MS 2003 Excel spreadsheet program and SPSS 12.0 package program. Menu mix% (MM%) and unit contribution margin were used as variables by Kasavana & Smith. Four possible classifications by Menu-Engineering technique were turned out as 'STAR', 'PLOWHORSE', 'PUZZLE', 'DOG'. The main menus served during a year were 128 dishes and about 141 peoples visited this restaurant daily. The mean age of the men was $44.1\;{\pm}\;6.3$, women were $32.7\;{\pm}\;6.4$ and showed that was statistically higher than that of women (p < .0001). The rates of STAR menus were 'Western style (75.0%)', 'guk/tang-ryu (48.1%)', 'jjigae/ jeongol-ryu (23.1%)', 'bap-ryu (17.2%)' in sequence. There were no STAR menus in gui/jorim/jjim-ryu. PLOWHORSE menus were 'gui-ryu (75.0%)', 'guk/tang-ryu (29.6%)', 'bap-ryu (27.6%)' in sequence. There were no PUZZLE or DOG menus in 'jjigae/jeongol-ryu'. PUZZLE menus were 'jorim/jjim-ryu and Myeonryu (each 33.3%)', 'bap-ryu (31.0%)' in sequence. PUZZLE menus were a lots of 'Chinese food (75.0%)' and 'myeonryu (55.6%)'. This study provides the basic data based on regularly menu analysis method applied the scientific menu analysis techniques in government offices food services, I'd like to suggest that the menu management must be done based on the necessity and result of menu analysis according to the seasonal and middle, long-term plans.

• Korean Chemical Engineering Research
• /
• v.58 no.3
• /
• pp.325-345
• /
• 2020

Determination of Best available technology (BAT) was suggested to reduce volatile organic compounds (VOCs) in a petrochemical industrial complex, by conducting human health risk, environmental, and economic assessment based on multimedia fugacity model. Fate and distribution of benzene, toluene, ethylbenzene, and xylene (BTEX) was predicted by the multimedia fugacity model, which represent VOCs emitted from the industrial complex in U-city. Media-integrated human health risk assessment and sensitivity analysis were conducted to predict the human health risk of BTEX and identify the critical variable which has adverse effects on human health. Besides, the environmental and economic assessment was conducted to determine the BAT for VOCs reduction. It is concluded that BTEX highly remained in soil media (60%, 61%, 64% and 63%), and xylene has remained as the highest proportion of BTEX in each environment media. From the candidates of BAT, the absorption was excluded due to its high human health risk. Moreover, it is identified that the half-life and exposure coefficient of each exposure route are highly correlated with human health risk by sensitivity analysis. In last, considering environmental and economic assessment, the regenerative thermal oxidation, the regenerative catalytic oxidation, the bio-filtration, the UV oxidation, and the activated carbon adsorption were determined as BAT for reducing VOCs in the petrochemical industrial complex. The suggested BAT determination methodology based on the media-integrated approach can contribute to the application of BAT into the workplace to efficiently manage the discharge facilities and operate an integrated environmental management system.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.15 no.4
• /
• pp.26-41
• /
• 2012

It is necessary to manage, forecast and prepare agricultural production based on accurate and up-to-date information in order to cope with the climate change and its impacts such as global warming, floods and droughts. This study examined the applicability as well as challenges of the object-based image analysis method for developing a land cover image classification algorithm, which can support the fast thematic mapping of wide agricultural areas on a regional scale. In order to test the applicability of RapidEye's multi-temporal spectral information for differentiating agricultural land cover types, the integration of other GIS data was minimized. Under this circumstance, the land cover classification accuracy at the study area of Kimje ($1300km^2$) was 80.3%. The geometric resolution of RapidEye, 6.5m showed the possibility to derive the spatial features of agricultural land use generally cultivated on a small scale in Korea. The object-based image analysis method can realize the expert knowledge in various ways during the classification process, so that the application of spectral image information can be optimized. An additional advantage is that the already developed classification algorithm can be stored, edited with variables in detail with regard to analytical purpose, and may be applied to other images as well as other regions. However, the segmentation process, which is fundamental for the object-based image classification, often cannot be explained quantitatively. Therefore, it is necessary to draw the best results based on expert's empirical and scientific knowledge.

• Journal of Plant Biotechnology
• /
• v.44 no.1
• /
• pp.97-106
• /
• 2017

Living modified crops are genetically modified living organisms and are widely used in biotechnical research and desired goods. As the reliance on LM products, concerns about safety of LMOs have been continuously increased in South Korea. We established the detection methods for unintentional released LMOs in environmental conditions. To detect six LM event genes of 1 canola, 1 maize and 4 soybeans, PCR conditions were based upon consideration of the Joint Research Centre information. Genomic DNAs were isolated from LM samples and PCR analysis were performed using each event-specific primer pair. Event-specific genes of all events were efficiently recognized by our methods. To investigate the insertion site of LM genes in each genome, we verified PCR product sequence by DNA sequencing. These results suggest that the LM event-specific gene amplification can be efficiently developed. In addition, our detection method is fit for monitoring and post-management of LM crops in the environment.

• Journal of Korea Water Resources Association
• /
• v.33 no.6
• /
• pp.745-755
• /
• 2000

The dam operation system of KOWACO for flood control doesn't have capability to account for the downstream hydrologic conditions and any feasible index to decide the pre-release from the forecasted rainfall and inflow. In this study, a dam operation model for flood control was developed to account for the flood flow condition of its downstream to give users the dam release schedules. Application test of EV ROM to Keum River showed that EV ROM is superior to the Rigid ROM and Technical ROM which are currently used by KOWACO. EV ROM developed in this study provides a release schedule accounting for the cumulative lateral flow hydrograph at the downstream control points where the discharge does not depend only on the dam operation. but also on lateral inflow from the tributaries. In order to reduce the peak discharge at the control points, it suggests the preliminary release during the early rising phase of the predicted hydrograph, holding the flood flow inside the dam during a peak phase, and afterward resuming the release. Three case studies of flood control by the operation of Daechung Multipurpose Dam in Geum River Basin show that the EV ROM is superior to the Rigid ROM and Technical ROM. This must be due to its nature to account for the downstream flow condition as well as the inflow and water level of the dam. It was also conceived that further case studies of EV ROM and more accurate rainfall prediction would improve the dam operation for flood control.ontrol.

• Journal of Food Hygiene and Safety
• /
• v.33 no.1
• /
• pp.23-30
• /
• 2018

The purpose of this research was to introduce network analysis method for analyzing pesticide residues in incongruity commercial agricultural commodities. Based on the "results in pesticide residues on incongruity commercial agricultural commodities" on "Guidelines for food safety management 2017", we used centrality analysis for pesticide residues via degree, closeness and betweenness centrality measurement. In case of degree centrality result, chlorpyrifos and diazinon were the most highly "connected node" in pesticide network. For the closeness centrality result, the most pesticides showed the similar closeness trend except for 19 species of pesticides. Fludioxonil and chlorpyrifos are recognized as the "bridge" of pesticides network with their high betweenness centrality. The results of network analysis show the "relation" data, which could not represent through out the conventional statistical analysis, among the pesticide residues. We hope that the network analysis method will be appropriate and precise tool for analyzing pesticide residues via elaboration and optimization.