• Journal of Korea Water Resources Association
• /
• v.56 no.6
• /
• pp.403-417
• /
• 2023

In this research, a methodology was developed for constructing an appropriate rainfall image database for estimating rainfall intensity based on CCTV video. The database was constructed in the Large-Scale Climate Environment Chamber of the Korea Conformity Laboratories, which can control variables with high irregularity and variability in real environments. 1,728 scenarios were designed under five different experimental conditions. 36 scenarios and a total of 97,200 frames were selected. Rain streaks were extracted using the k-nearest neighbor algorithm by calculating the difference between each image and the background. To prevent overfitting, data with pixel values greater than set threshold, compared to the average pixel value for each image, were selected. The area with maximum pixel variability was determined by shifting with every 10 pixels and set as a representative area (180×180) for the original image. After re-transforming to 120×120 size as an input data for convolutional neural networks model, image augmentation was progressed under unified shooting conditions. 92% of the data showed within the 10% absolute range of PBIAS. It is clear that the final results in this study have the potential to enhance the accuracy and efficacy of existing real-world CCTV systems with transfer learning.

• Journal of Environmental Science International
• /
• v.15 no.12
• /
• pp.1171-1175
• /
• 2006

Deposited road particles (DRPs) were analysed for heavy metal concentrations at four different roads in a city, Korea. The samples were collected using a roadway surface vacuum cleaning vehicle which was commonly used in collecting roadway surface particles. Six particle size ranges were analyzed separately for twelve heavy metal elements (Cd, Cr, Pb, Ni, Al, As, Co, Cu, Fe, Mn, Zn and Hg). At all sampling sites, the high concentration of the heavy metals occurred in the <74um particle size range, which conventional roadway cleaning vehicles do not remove efficiently. The Pb concentration significantly increased with decreasing particle size of DRPs, and other toxic heavy metals (Cd, Cr and Ni) also showed similar results. The heavy metal concentrations in the smaller size fraction of DRPs is important because they are contaminants that are preferentially transported by road runoff during rainfall.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2005.05b
• /
• pp.123-126
• /
• 2005

본 연구에서는 상층기상자료, 자동 기상 관측망 자료 및 신경망기법을 사용하여 단시간 강우 예측 모형을 개발하였다. 호우를 동반한 이송 기상 시스템의 이동 경로가 라디오존데로부터 획득할 수 있는 상층기상 자료 즉 상층 풍향자료와 동일한 방향으로 이동한다는 가정 하에 원거리에서 발생하는 기상현상의 발달과정을 판단 할 수 있는 알고리즘을 개발하고, 이러한 원거리 입력 자료와 예측하고자 하는 값 사이의 비선형 상관 관계를 연결하는 기법으로 인공 신경망 기법을 도입하였다. 개발된 모형을 2002년 태풍 루사로 인하여 큰 피해를 입은 감천지역에 적용하였다. 포항과 오산의 라디오존데에서 획득한 700mb에서의 풍향자료와 5년의 자료기간을 가지는 350개의 자동 기상 관측망 자료를 입력 자료로 사용하였으며 결과는 상층기상자료를 사용하지 않고 예측한 결과에 대하여 개선된 강우 예측결과를 보여주었다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2006.05a
• /
• pp.623-626
• /
• 2006

일정한 시간간격으로 제공되는 연속된 두 장의 레이더 반사도(합성 CAPPI) 자료의 최대 상관계수를 찾아 냄으로써 강수의 움직임을 산출하는 TREC(Tracking Radar Echoes by Correlation) 기법은 동일 고도의 레이더 반사도 자료를 이용하기 때문에 수평방향의 2차원이며, 대류성 구름체계에서 발생되는 수직 활동을 표현할 수 없는 한계성을 지니고 있다. 본 연구에서는 여러 고도의 레이더 반사도 자료를 이용하여 기존의 TREC 기법을 이용한 단시간 예보모형을 개선하고자 하였다. 특정고도의 레이더 반사도를 이용하여 에코를 추적하는 TREC 기법의 단점을 보완하기 위하여 서로 다른 고도의 레이더 반사도를 이용함으로써 기존의 접근법보다 실제 강수의 움직임에 더욱 가깝도록 단시간 강우 예보 정확도를 개선하였다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2006.05a
• /
• pp.1470-1473
• /
• 2006

연속된 두 장의 레이더 반사도(합성 CAPPI)를 이용하여 설정된 두 윈도우 사이의 최대 상관계수를 찾아 냄으로써 강수의 움직임을 파악하는 기존의 TREC(Tracking Radar Echoes by Correlation) 기법은 단지 통계적인 상관법을 이용하여 산출된 TREC 벡터를 외삽하기 때문에 강우 시스템의 이동양상을 물리적으로 표현하는데 한계를 가질 뿐만 아니라 강수가 직선운동을 하는 것처럼 묘사될 수밖에 없는 기법의 한계성을지니고 있다. 본 연구에서는 도플러 레이더로부터 생산되는 시선속도를 이용하여 바람장을 산출하고 이를 TREC 벡터와 연계시켜 단시간 예보모형을 개선하고자 하였다. 시선속도는 레이더로부터 멀어지거나 다가오는 물체의 속도성분이며, 이를 이용하여 강수 영역 내의 바람장을 산출할 수 있다. 이러한 바람장 정보와 연계한 TREC 벡터의 개선은 단시간 강우 예보모형의 개선을 통하여 짧은 시간에 급격한 발달하는 집중호우 등에 대한 보다 정확한 예보를 가능하게 한다.

• 한국지구물리탐사학회:학술대회논문집
• /
• 2003.11a
• /
• pp.703-710
• /
• 2003

In the sub-surface environments, detection of the movement of contaminant substances and recharge of groundwater by rainfall are very important factors which contain porosity and effective porosity of porous media. In this paper, the applicability of permittivity methods and proposed dielectric mixing models (DDMs) are discussed. This study showed that the ratio of effective porosity to porosity of Toyoura and River sands were 0.856 and 0.843. From the relationships between the relative porosity and effective porosity, all measured values can be confirmed to outside the range to about 0.800 for Toyoura and River sands under all experiments by FDR and FDR-V systems. In the study, this permittivity equipment can be considered to be good enough to measure determining the physical parameters of saturated soils. Consequently, this permittivity method can be contributed to estimate a porosity and effective porosity of saturated porous media because it is easy and instantaneous than previous in-situ methods.

• International Journal of Computer Science & Network Security
• /
• v.23 no.1
• /
• pp.33-40
• /
• 2023

Machine-learning techniques are discovering effective performance on data analytics. Classification and regression are supported for prediction on different kinds of data. There are various breeds of classification techniques are using based on nature of data. Threshold determination is essential to making better model for unlabelled data. In this paper, threshold value applied as range, based on min-max normalization technique for creating labels and multiclass classification performed on rainfall data. Binary classification is applied on autism data and classification techniques applied on child abuse data. Performance of each technique analysed with the evaluation metrics.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.59 no.2
• /
• pp.1-11
• /
• 2017

The analysis of baseflow contribution is very significant in Korea because most rivers have high variability of streamflow due to the monsoon climate. Recently, the importance of such analysis is being more evident especially in terms of river management because of the changing pattern of rainfall and runoff resulted from climate change. Various baseflow separation methods have been developed to separate baseflow from streamflow. However, it is very difficult to identify which method is the most accurate way due to the lack of measured baseflow data. Moreover, it is inappropriate to analyze the annual baseflow contribution for Korean rivers because rainfall patterns varies significantly with the seasons. Thus, this study compared the baseflow contributions at various time-scales (annual, seasonal and monthly) for the 4 major river basins through BFI (baseflow index) and suggested baseflow contribution of each basin by the BFI ranges searched from different baseflow separation methods (e.g., BFLOW, HYSEP, PART, WHAT). Based on the comparison of baseflow contributions at the three time scales, this study showed that the baseflow contributions from the monthly and seasonal analysis are more reasonable than that from the annual analysis. Furthermore, this study proposes that defining BFI with its range is more proper than a specific value for a watershed, considering the difference of BFIs between various baseflow separation methods.

• Environmental Engineering Research
• /
• v.18 no.2
• /
• pp.95-101
• /
• 2013

Alternative sources of water are sought in some water stressed communities in the study area. The study focuses on the Aguata-Awka area of southeastern Nigeria. Aquifers occur at great depths, and surface waters may be far from homesteads. The scarcity of water has necessitated the people to adopt various local technologies for harvesting rainfall. The local technology includes collecting rainwater from roofs and channeling the water into large underground tanks, shallow wells and surface reservoirs. Large concrete tanks of $6m{\times}6m{\times}4m$ dimensions are often built underground and can store $144m^3$ of water. Surface reservoirs built on 4 m concrete pillar supports having dimensions of $10m{\times}10m{\times}4m$ and have a storage capacity of $400m^3$. Water samples were collected at 3 different locations of Agulu, Ekwulobia, and Awka and were analyzed for their physical, chemical, and bacteriological parameters. Results indicate a range of values for pH, 5.9 to 7.1; turbidity, 0.9 to 2.7; total dissolved solids, 80 to 170 mg/L; total hardness, 4.5 to 6.4 mg/L; magnesium, 1.2 to 1.4 mg/L; bicarbonate, 19.4 to 83.6 mg/L; and sulfate, 3.6 to 6.4 mg/L. Bacteriological analysis results were negative for fecal and total coliform counts. All parameters, with the exception of pH where aluminum and galvanized iron roofs are used for collection, fall within the recommended guidelines for drinking water quality of the World Health Organization, and the Standard Organization of Nigeria, new Nigerian standards for drinking water quality. Magnesium is above the maximum permitted level for consumer acceptability of the Nigerian standards for drinking water quality. The water can be classified as fresh moderately hard and soft. The water can be described as a calcium and bicarbonate type.

• Journal of Korea Water Resources Association
• /
• v.37 no.1
• /
• pp.67-75
• /
• 2004

In this study, real-time forecasting model(Neural Dam Inflow Forecasting Model; NDIFM) based on neural network to predict the dam inflow which is occurred by flood runoff is developed and applied to check its availability for the operation of multi-purpose reservoir Developed model Is applied to predict the flood Inflow on dam Nam-Gang in Nak-dong river basin where the rate of flood control dependent on reservoir operation is high. The input data for this model are average rainfall data composed of mean areal rainfall of upstream basin from dam location, observed inflow data, and predicted inflow data. As a result of the simulation for flood inflow forecasting, it is found that NDIFM-I is the best predictive model for real-time operation. In addition, the results of forecasting used on NDIFM-II and NDIFM-III are not bad and these models showed wide range of applicability for real-time forecasting. Consequently, if the quality of observed hydrological data is improved, it is expected that the neural network model which is black-box model can be utilized for real-time flood forecasting rather than conceptual models of which physical parameter is complex.