• Journal of Korea Water Resources Association
• /
• v.54 no.5
• /
• pp.301-309
• /
• 2021

In this study, we tried to improve the performance of the existing U-net-based deep learning rainfall prediction model, which can weaken the meaning of time series order. For this, ConvLSTM2D U-Net structure model considering temporal consistency of data was applied, and we evaluated accuracy of the ConvLSTM2D U-Net model using a RainNet model and an extrapolation-based advection model. In addition, we tried to improve the uncertainty in the model training process by performing learning not only with a single model but also with 10 ensemble models. The trained neural network rainfall prediction model was optimized to generate 10-minute advance prediction data using four consecutive data of the past 30 minutes from the present. The results of deep learning rainfall prediction models are difficult to identify schematically distinct differences, but with ConvLSTM2D U-Net, the magnitude of the prediction error is the smallest and the location of rainfall is relatively accurate. In particular, the ensemble ConvLSTM2D U-Net showed high CSI, low MAE, and a narrow error range, and predicted rainfall more accurately and stable prediction performance than other models. However, the prediction performance for a specific point was very low compared to the prediction performance for the entire area, and the deep learning rainfall prediction model also had limitations. Through this study, it was confirmed that the ConvLSTM2D U-Net neural network structure to account for the change of time could increase the prediction accuracy, but there is still a limitation of the convolution deep neural network model due to spatial smoothing in the strong rainfall region or detailed rainfall prediction.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.6
• /
• pp.1515-1521
• /
• 1996

Object-oriented and distributed processing methods are adopted next -generation telecommunications management architecture. In this paper, we design and implement gradual federation function on connection management system providing connection services to transport network for interworking and encapsulated trader and gateway function for testing interoperability of these service objects between distributed processing environments. These techniques are practically applicable to support interworking between heterogeneous management network systems according to a tendency of integrating telecommunications management systems.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.18 no.4
• /
• pp.131-144
• /
• 2015

This research proposes a prototype for visualizing radar rainfall data using the unity 3D engine. The mashup of radar data with topographic information is necessary for the 3D visualization of the radar data with high quality. However, the mashup of a huge amount of radar data and topographic data causes the overload of data processing and low quality of the visualization results. This research utilized the Unitiy 3D engine, a widely used engine in the game industry, for visualizing the 3D topographic data such as the satellite imagery/the DEM(Digital Elevation Model) and radar rainfall data. The satellite image segmentation technique and the image texture layer mashup technique are employed to construct the 3D visualization system prototype based on the topographic information. The developed protype will be applied to the disaster-prevention works by providing the radar rainfall data with the 3D visualization based on the topographic information.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.44 no.3
• /
• pp.385-396
• /
• 2024

To enhance the accuracy of sinkhole detection using GPR, this study derived a convolutional neural network that can optimally extract sinkhole characteristics from GPR B-scan grayscale images. The pre-trained convolutional neural network is evaluated to be more than twice as effective as the vanilla convolutional neural network. In pre-trained convolutional neural networks, fast feature extraction is found to cause less overfitting than feature extraction. It is analyzed that the top-1 verification accuracy and computation time are different depending on the type of architecture and simulation conditions. Among the pre-trained convolutional neural networks, InceptionV3 are evaluated as most robust for sinkhole detection in GPR B-scan grayscale images. When considering both top-1 verification accuracy and architecture efficiency index, VGG19 and VGG16 are analyzed to have high efficiency as the backbone for extracting sinkhole feature from GPR B-scan grayscale images. MobileNetV3-Large backbone is found to be suitable when mounted on GPR equipment to extract sinkhole feature in real time.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.97-97
• /
• 2023

최근 재해분야에서 인공신경망(ANN), 기계학습(ML), 딥러닝(DL) 등 AI 기술이 활용성이 점차 증가하고 있으며, 센싱정보와 연계한 시설물 안전관리, 원격탐사와 연계한 재해감시(녹조, 산사태, 산불 등), 수문시계열(수위, 유량 등) 예측, 레이더·위성강수 자료의 보정과 예측, 상하수도 관망누수예측 등 다양한 분야에서 AI 기술이 적용되고 그 활용성이 검증된 바 있다. 본 연구에서는 ML, DL, 물리기반신경망(Pysics-informed Neural Networks, PINNs)을 이용한 다양한 재해분석 사례를 소개하고, 그 활용성과 한계에 대해서 논의하고자 한다. 주요사례로는 (1) SAR영상과 기계학습을 이용한 재해피해지역(울진 산불) 감지, (2) 국가 디지털 정보를 이용한 산사태 위험지역 판별(인제 산사태) (3) 기계학습 및 딥러닝 기법을 이용한 위성강수 자료의 보정·예측 및 유출해석, (4) 수리해석을 위한 수치해석분야에서의 PINNs의 적용성(1차원 Saint-Venant 식 해석) 평가 연구결과를 공유한다. 특히, 자료의 입·출력 자료만으로 학습된 인공신경망 모형 대신 지배방정식(물리방정식)을 만족하도록 강제한 PINNs의 경우, 인공신경망 모형보다 우수한 모의능력을 보여주었으며, 향후 복잡한 수리모델링 등 수치해석분야에서 그 활용가능성이 매우 높을 것으로 판단된다.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2003.05a
• /
• pp.87-91
• /
• 2003

It has well known that MUSIC and ESPRIT algorithms estimate angle of arrival(AOA) with high resolution by eigenvalue decomposition of the covariance matrix which were obtained from the array antennas. However, the disadvantage of MUSIC and ESPRIT is that they are computationally ineffective, and then they are difficult to implement in real time. The other problem of MUSIC and ESRPIT is to require calibrated antennas with uniform features, and are sensitive to the manufacturing facult and other physical uncertainties. To overcome these disadvantages, several method using neural model have been study. For multiple signals, those require huge training data prior to AOA estimation. This paper proposes the algorithm for AOA estimation by interconnected hopfield neural model. Computer simulations show the validity of the proposed algorithm. The proposed method does not require huge training procedure and only assigns interconnected coefficients to the neural network prior to AOA estimation.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.19 no.4
• /
• pp.443-452
• /
• 2016

Current low-altitude radar system often fails to detect small unmanned aerial vehicles (UAV) because of their small radar cross section (RCS) compared with larger targets. As a potential alternative, a passive bistatic radar system has been considered. We study an optimal deployment problem for the passive bistatic radar system. We model this problem as a covering problem, and develop an integer programming model. The objective of the model is to maximize coverage of a passive bistatic radar system. Our model takes into account factors specific to a bistatic radar system, including bistatic RCS and transmitter-receiver pair coverage. Considering bistatic RCS instead of constant RCS is important because the slight difference of RCS value for small UAVs could significantly influence the detection probability. The paired radar coverage is defined by using the concept of gradual coverage and cooperative coverage to represent a realistic environment.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.21 no.2
• /
• pp.133-140
• /
• 2018

In electronic warfare(EW), low probability of intercept(LPI) radar signal is a survival technique. Accordingly, identification techniques of the LPI radar waveform have became significant recently. In this paper, classification and extracting parameters techniques for 7 intrapulse modulated radar signals are introduced. We propose a technique of classifying intrapulse modulated radar signals using Convolutional Neural Network(CNN). The time-frequency image(TFI) obtained from Choi-William Distribution(CWD) is used as the input of CNN without extracting the extra feature of each intrapulse modulated radar signals. In addition a method to extract the intrapulse radar modulation parameters using binary image processing is introduced. We demonstrate the performance of the proposed intrapulse radar waveform identification system. Simulation results show that the classification system achieves a overall correct classification success rate of 90 % or better at SNR = -6 dB and the parameter extraction system has an overall error of less than 10 % at SNR of less than -4 dB.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.132-132
• /
• 2022

강우레이더는 넓은 공간에서의 조밀한 정보를 제공하여 돌발홍수 정보 제공에 많은 장점을 보유하고 있다. 이에 한국건설기술연구원은 강우레이더의 장점을 활용하여 행정동(읍면동) 단위로 3시간 전 3단계 돌발홍수 예측 정보(주의/경계/심각)를 제공하는 돌발홍수예측 시스템을 구축하였다. 행정동 단위의 돌발홍수 예측 정보가 제공됨으로써 기존 국가 하천 중심의 홍수 예보 시스템에서 제공되지 못했던 홍수예보 취약지역에서의 홍수 예보가 가능해졌다. 하지만 돌발홍수예측 시스템에서 제공되고 있는 돌발홍수 예측 정보의 신뢰성을 높이기 위해서는 제공되고 있는 정보의 정확도가 확보 되어야 한다. 이에 본 연구에서는 돌발홍수 예측 정보의 실증을 위해 낙동강홍수통제소 유역 내에서의 홍수예보 취약지역 지점을 선정하였다. 취약지역은 도심지, 산지·소하천, 해안지역으로 구분하여 선정되었다. 또한 돌발홍수예측 시스템 내에서의 돌발홍수위험 기준은 전국 피해사례에 대한 통계적으로 추정한 값으로, 실제 홍수취약 지역에서의 위험 기준과 다소 차이가 나타날 수 있다. 이에 본 연구에서는 선정된 시범 지역에서의 돌발홍수 위험 기준을 추정하기 위해 시범 지역에서 발생한 강우 특성을 분석하였다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2016.05a
• /
• pp.287-287
• /
• 2016

최근 기후변화로 인해 돌발성 집중호우가 증가하는 추세로 홍수피해가 발생하고 있는데 이러한 피해를 예방하고 빠른 대처를 위해 강우의 정밀한 관측뿐만 아니라 강우의 정확한 공간 분포 파악에 대한 필요성이 중요하게 대두되고 있다. 그러나 일반적으로 지상우량계의 경우, 공간적인 강우분포 분석에 한계가 존재하여 레이더 강우자료와 함께 활용하는 연구가 진행되어 왔다. 따라서 본 연구에서는 강우장 추정시, 공간보간 기법인 크리깅 기법을 적용하여 강우장을 추정하고 유출 해석을 통해 그 적용성을 확인하고자 하였다. 국내에서 일반적으로 사용되는 크리깅 기법인 OK(Ordinary Kriging), CK(Co-Kriging) 외에도 KED(Kriging with External Drift) 기법을 적용하여 강우장을 추정하고 분포형 수문모형인 $Vflo^{TM}$의 입력자료로 사용하여 유출해석시 정확도를 비교 분석하였다. 추정된 강우장의 정량적 평가 결과, 지상강우만을 이용하는 OK 기법이 가장 우수한 결과를 나타내었다. 하지만 강우의 공간 분포 특성 반영 측면에서는 KED와 CK가 보다 더 좋은 결과를 나타내었다. 또한 유출해석의 경우 지형학적 매개변수 조정에 의한 강우 입력자료의 왜곡을 배제하기 위해 검 보정은 실시하지 않았으며 오차분석 결과에서 KED, CK, OK, Radar 순으로 관측유량을 잘 재현하는 것으로 확인되었다. 본 연구를 통해 공간보간 기법의 수문학적 적용성을 확인하였으며 모형의 검 보정을 통해 수문모형의 입력자료로서 활용성을 가질 수 있을 것으로 판단된다. 또한 이를 통해 생성된 강우장을 활용한다면, 관측망의 밀도가 낮은 지역과 미계측 유역 등에 적용하여 수문시스템해석에 도움이 될 것으로 판단된다.