• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.344-344
• /
• 2023

최근 기후변화에 따라 극한 강우로 전 세계적으로 국지적 홍수 피해가 증가하고 있다. 또한 극한 강우 발생시 다양한 건설 현장의 상황에 따라 침수 취약성이 나타나 인적 물적 피해로 이어질 수 있다. 특히, 시공에 따른 현장 지형 변화에 대해 실시간으로 침수 예측이 불가하여 위험 판단이 어려운 실정이며, 극한 강우 발생에 대비하기 위해 강우 정보 획득 및 분석을 효율화하여 강우예측 정확성을 높일 필요가 있다. 이러한 필요성에 따라 본 연구에서는 건설 현장의 침수 피해를 최소화하기 위해 침수 위험을 판정하고 예측하는 방법을 제시하고자 한다. 본 연구의 침수 위험 판정 방법은 건설 현장에서 실시간 지형변화 정보 확보와 침수 위험 판정의 정확도를 높이기 위한 침수심 분석에 인공 신경망 기법을 활용하였다. 또한, 침수판정 알고리즘은 지형, 강우 분석 모듈과 침수판정 모듈로 구성하였다. 지형 분석 모듈은 건설 현장이 시공진행에 따른 지형 데이터의 변화를 고려하기 위해 실시간 영상 정보의 객체 탐지를 구분하는 인공 신경망 기법을 적용해 지형 분석 모듈을 구축하였다. 강우 분석 모듈은 다양한 강우 정보를 취합할 수 있는 서버를 구축하여 강우 임베딩 정보를 실시간으로 분석하도록 고안하여 정확도를 높였다. 이러한 자료를 바탕으로 강우-유출해석에 의한 침수심 값과 실측값, 침수 지표를 활용하여 인공 신경망 기법으로 침수 위험을 판정하도록 제시하였다. 본 연구를 통해 건설 현장에서 지형 상태의 지속적인 변화와 강우데이터의 정확도 향상에 대응할 수 있는 침수 위험 판정이 가능하고 인적 물적 피해 최소화를 기대할 수 있다. 향후, 본 연구에서 제시된 방법은 건설 현장에서 분석 시스템과 실측 모니터링에 의해 검증되어야 할 것이며, 건설 현장 외에도 스마트 도시 및 지하 공간에서 확대하여 적용할 수 있을 것으로 판단된다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.417-417
• /
• 2023

도시홍수는 도시의 주요 기능을 마비시킬 수 있는 수재해로서, 최근 집중호우로 인해 홍수 및 침수 위험도가 증가하고 있다. 집중호우는 한정된 지역에 단시간 동안 집중적으로 폭우가 발생하는 현상을 의미하며, 도시 지역에서 강우 추정 및 예보를 위해 레이더의 활용이 증대되고 있다. 레이더는 수상체 또는 구름으로부터 반사되는 신호를 분석해서 강우량을 측정하는 장비이다. 기상청의 기상레이더(S밴드)의 주요 목적은 남한에 발생하는 기상현상 탐지 및 악기상 대비이다. 관측반경이 넓기에 도시 지역에 적합하지 않는 반면, X밴드 이중편파레이더는 높은 시공간 해상도를 갖는 관측자료를 제공하기에 도시 지역에 대한 강우 추정 및 예보의 정확도가 상대적으로 높다. 따라서, 본 연구에서는 딥러닝 기반 초해상화(Super Resolution) 기술을 활용하여 저해상도(Low Resolution. LR) 영상인 S밴드 레이더 자료로부터 고해상도(High Resolution, HR) 영상을 생성하는 기술을 개발하였다. 초해상도 연구는 Nearest Neighbor, Bicubic과 같은 간단한 보간법(interpolation)에서 시작하여, 최근 딥러닝 기반의 초해상화 알고리즘은 가장 일반화된 합성곱 신경망(CNN)을 통해 연구가 이루어지고 있다. X밴드 레이더 반사도 자료를 고해상도(HR), S밴드 레이더 반사도 자료를 저해상도(LR) 입력자료로 사용하여 초해상화 모형을 구성하였다. 2018~2020년에 발생한 서울시 호우 사례를 중심으로 데이터를 구축하였다. 구축된 데이터로부터 훈련된 초해상도 심층신경망 모형으로부터 저해상도 이미지를 고해상도로 변환한 결과를 PSNR(Peak Signal-to-noise Ratio), SSIM(Structural SIMilarity)와 같은 평가지표로 결과를 평가하였다. 본 연구를 통해 기존 방법들에 비해 높은 공간적 해상도를 갖는 레이더 자료를 생산할 수 있을 것으로 기대된다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.12
• /
• pp.1872-1879
• /
• 2022

Recently, with the development of artificial intelligence technology, research to use artificial intelligence to detect hacking attacks is being actively conducted. However, the fact that security data is a representative imbalanced data is recognized as a major obstacle in composing the learning data, which is the key to the development of artificial intelligence models. Therefore, in this paper, we propose a W-VAE oversampling technique that applies VAE, a deep learning generation model, to data extraction for oversampling, and sets the number of oversampling for each class through weight calculation using K-NN for sampling. In this paper, a total of five oversampling techniques such as ROS, SMOTE, and ADASYN were applied through NSL-KDD, an open network security dataset. The oversampling method proposed in this paper proved to be the most effective sampling method compared to the existing oversampling method through the F1-Score evaluation index.

• Journal of Advanced Navigation Technology
• /
• v.25 no.6
• /
• pp.476-483
• /
• 2021

In order to maintain the safety of the airspace with ever increasing traffic volume, it is necessary to thoroughly analyze the collision risk with the current data. In this study, collision risk analysis was conducted using Detect and Avoid (DAA) Well-Clear (DWC) metrics, risk induces developed for the DAA systems of unmanned aerial vehicles. All flights in year 2019 that flew within the Incheon Flight Information Region (FIR) boundary were analyzed using the recorded Automatic Dependent Surveillance-Broadcast(ADS-B) data. High risk regions as well as trends by airports and seasons were identified. The results indicate that the risk is higher around the congested area near Incheon International Airport and Gimpo International Airport. Seasonally, the risk was highest in August that coincides with the Summer vacation period. The result will be useful for the baseline data for various aviation safety enhancement activities such as revision of routes and procedures and training of the field specialists.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.1
• /
• pp.59-66
• /
• 2024

In this paper, the method to design the antenna, which is suitable for an impulse-like waveform radiation, is presented. In general, the impulse-like waveform has its spectrum of around sub GHz bandwidth and the antenna should be properly designed for not only operating wide-bandwidth also reflecting the time domain characteristics for near-zone impulse radar applications. In this regard, Vivaldi antenna has been designed and characterized in terms of short-pulse radiating aspects in the time domain and verified by measured results. The designed antenna has shown to be operating within wide-bandwidth and to be stable for the input impedance from 1.8 to more than 10GHz. The far-zone radiating waveform has been investigated on each plane at the interval of 30degree and the designed antenna has shown to be a directive characteristic. It can be seen that those results proposed are widely applicable to the near area sensing applications such as ground-penetrating radar.

• Korean Journal of Remote Sensing
• /
• v.39 no.6_1
• /
• pp.1413-1425
• /
• 2023

The increasing frequency of wildfires due to climate change is causing extreme loss of life and property. They cause loss of vegetation and affect ecosystem changes depending on their intensity and occurrence. Ecosystem changes, in turn, affect wildfire occurrence, causing secondary damage. Thus, accurate estimation of the areas affected by wildfires is fundamental. Satellite remote sensing is used for forest fire detection because it can rapidly acquire topographic and meteorological information about the affected area after forest fires. In addition, deep learning algorithms such as convolutional neural networks (CNN) and transformer models show high performance for more accurate monitoring of fire-burnt regions. To date, the application of deep learning models has been limited, and there is a scarcity of reports providing quantitative performance evaluations for practical field utilization. Hence, this study emphasizes a comparative analysis, exploring performance enhancements achieved through both model selection and data design. This study examined deep learning models for detecting wildfire-damaged areas using Landsat 8 satellite images in California. Also, we conducted a comprehensive comparison and analysis of the detection performance of multiple models, such as U-Net and High-Resolution Network-Object Contextual Representation (HRNet-OCR). Wildfire-related spectral indices such as normalized difference vegetation index (NDVI) and normalized burn ratio (NBR) were used as input channels for the deep learning models to reflect the degree of vegetation cover and surface moisture content. As a result, the mean intersection over union (mIoU) was 0.831 for U-Net and 0.848 for HRNet-OCR, showing high segmentation performance. The inclusion of spectral indices alongside the base wavelength bands resulted in increased metric values for all combinations, affirming that the augmentation of input data with spectral indices contributes to the refinement of pixels. This study can be applied to other satellite images to build a recovery strategy for fire-burnt areas.

• Journal of Bio-Environment Control
• /
• v.28 no.4
• /
• pp.461-470
• /
• 2019

This study was carried out to quantify the drought stress in grafted watermelon seedlings non-destructively by using chlorophyll fluorescence (CF) imaging technique rather than the visual judgment. Six-day old watermelon seedlings were grown under uniform irrigation for 3 days, and then given drought stress. Afterward, the sensor for the measurement of water content in plug tray cell unit was used to classify the drought-stress level into nine groups from D1 (53.0%, sufficient moisture state) to D9 (15.7%, extremely dry stress), and the 16 CF parameters were measured. In addition, re-irrigation was performed on the drought stressed seedlings(D5 - D9) to determine the growth and photosynthesis recovery level, which was not confirmed by visual judgment. The kinetic curve patterns of CF in three different drought stressed seedling groups were found to be different for the early detection of drought stress. All the 16 CF parameters decreased continuously with exposure to drought stress and drastically decreased from D5 (32.1%) where the visual judgment was possible. The fluorescence decline ratio (Rfd_Lss) started to decrease from the initial drought stress level (D5 - D6), and the Maximum PSII quantum yield (Fv/Fm) was significantly decreased in the later extreme drought stress range (D7 - D9) by re-irrigation recovery test. Thus, Rfd_Lss and Fv/Fm parameters were finally selected as potent indicators of growth and photosynthesis recovery in the initial and later stages of drought stress. Also, to the differences in the numerical values of the individual chlorophyll fluorescence parameters, the drought stress level was intuitively confirmed through the image. These results indicate that Rfd and Fv/Fm can be considered as potential CF parameters for the detection of low and extremely high drought stress, respectively. Furthermore, Fv/Fm can be considered as the best CF parameters for recovery at re-irrigation.

• Geophysics and Geophysical Exploration
• /
• v.8 no.1
• /
• pp.50-58
• /
• 2005

Recently, the imaging of geological structures beneath water-covered areas has been in great demand because of numerous tunnel and bridge construction projects on river or lake sites. An electrical resistivity survey can be effective in such a situation because it provides a subsurface image of faults or weak zones beneath the water layer. Even though conventional resistivity surveys in water-covered areas, in which electrodes are installed on the water bottom, do give high-resolution subsurface images, much time and effort is required to install electrodes. Therefore, an easier and more convenient method is sought to find the strike direction of the main zones of weakness, especially for reconnaissance surveys. In this paper, we investigate the applicability of the streamer resistivity survey method, which uses electrodes in a streamer cable towed by ship or boat, for delineating a fault zone. We do this through numerical experiments with models of water-covered areas. We demonstrate that the fault zone can be imaged, not only by installing electrodes on the water bottom, but also by using floating electrodes, when the depth of water is less than twice the electrode spacing. In addition, we compare the signal-to-noise ratio and resolving power of four kinds of electrode arrays that can be adapted to the streamer resistivity method. Following this numerical study, we carried out both conventional and streamer resistivity surveys for the planned tunnel construction site located at the Han River in Seoul, Korea. To obtain high-resolution resistivity images we used the conventional method, and installed electrodes on the water bottom along the planned route of the tunnel beneath the river. Applying a two-dimensional inversion scheme to the measured data, we found three distinctive low-resistivity anomalies, which we interpreted as associated with fault zones. To determine the strike direction of these three fault zones, we used the quick and convenient streamer resistivity.

• Journal of the Korean Applied Science and Technology
• /
• v.36 no.3
• /
• pp.1028-1040
• /
• 2019

The current study was performed to investigate the magnitude of exercise-induced muscle damage (EIMD) after downhill running (DR) of different intensities and to examine the availability of muscle echo intensity as biomarkers to detect regional damage within quadriceps muscle group (QG) following DR. Healthy college-age men (n=11) were experienced twice DR sessions [$50%HR_{max}$ DR, LDR; $70%HR_{max}$ DR, HDR] separated by a 2-week wash-out period with the random order. After DR, severity of EIMD according to exercise intensity were determined by serum creatine kinase (CK) activity, muscle tenderness, and neuromuscular function indicators such as a maximal voluntary isometric contraction (MVIC) and range of motion (ROM). Transvaginal B-mode imaging had been employed to evaluate regional muscle echo intensity within QG [rectus femoris, RF; vastus lateralis, VL; vastus medialis, VM; vastus intermedius, VI]. After both DR sessions, changes in serum CK activity and muscle tenderness have tended to more increase in HDR compared to those of LDR. There was a significant interaction effect between exercise intensity during DR and the time course of serum CK activity(p<.05). However, there were no statistical differences between sessions in muscle tenderness. The time course of changes in the neuromuscular functions after DR were similar to those of regional muscle echo intensity regardless exercise intensity. Although neuromuscular function showed to decline in HDR more than those of LDR after DR, no statistical differences between sessions. In contrast, there were significant interaction effects between sessions and time course of changes in RF and VL muscle echo intensity(p<.01), but not shown in those of VI and VM. These results indicated that each muscles within the QG show different response profiles for EIMD during DR, exercise intensity influences on these responses as well. In particular, current findings suggested that muscle echo intensity derived from ultrasound imaging is capable of detecting regional muscle damage in QG following DR.

• Geophysics and Geophysical Exploration
• /
• v.11 no.4
• /
• pp.368-372
• /
• 2008

This study evaluates the usefulness and capability of surface electrical resistivity technique for identifying the weak zones or subsurface cavities in karst area with limestone formation. Weak zones or cavities near surface can be potentially dangerous and several problems are associated with collapse of roads or buildings accompanied by subsidence phenomena. In this paper, both two and three dimensional resistivity investigation were conducted to investigate subsidence along a road in Yongweol-ri, Muan-gun, South Korea. The results of the resistivity survey using dipole-dipole array provide a clear view of the weathered regolith, the distribution of weak zones or cavities and bedrock. Several low resistivity areas were identified and subsequent drilling led to the discovery of several weak zone or clay-filled underground cavities. The drilling results show excellent correlation with the resistivity images. It is illustrated, the ability of electrical technique to produce high resolution images of subsurface, which are useful for subsidence assessment. Also the results of this study have demonstrated that two and three dimensional electrical resistivity surveys are useful for delineating the subsidence area. Based on resistivity imaging, the map of hazardous zone has been developed.