• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.137-137
• /
• 2023

해양쓰레기는 해안의 심미적 가치 저하뿐만 아니라 생태계 파괴, 유령 어업에 따른 수산업 피해 등의 사회적·환경적 문제를 발생시키며, 그중 70% 이상은 육상 기인으로 플라스틱 및 기타 쓰레기가 주를 이루는 해외와 달리 국내의 경우 다량의 초목류를 포함하고 있다. 다양한 부유쓰레기에 대한 기존의 해양쓰레기량 추정의 한계와 하천·하구 쓰레기 수거의 효율화를 위해 해양으로 유입되는 부유쓰레기 방지를 위한 실효성 있는 대책 수립이 필요한 실정이다. 본 연구는 해양 유입 전 하천의 차단시설에 차집된 부유쓰레기의 수거 효율화 및 지속가능한 해양쓰레기 데이터 구축을 위해 AI기반의 기술을 통해 부유쓰레기 성상 분석 기법(Object Detection)과 차집량 분석 기법(Semantic Segmentation)을 활용하였다. 실제와 유사한 데이터 수집을 위해 다양한 하천 환경(정수조, 소하천, 급경사수로)에 대해 탁도(녹조, 유사), 광량, 쓰레기형상, 초목류 함량, 날씨(소하천), 유속(급경사수로) 등의 실험조건에 대하여 해양쓰레기 분류 기준 및 통계를 바탕으로 부유쓰레기 종류 선정하여 학습을 위한 데이터를 수집하였다. 학습 목적에 따라 구분하여 라벨링(Bounding box, Polygon)을 수행하고, 각 분석 기법별 전이학습을 통해 Phase 1(정수조), Phase 2(소하천), Phase 3(급경사수로) 순서로 모델을 고도화하였다. 성상 분석을 위해 YOLO v4를 활용하여 Train, Test DataSet(9:1)을 구성하고 학습 및 평가는 Iteration마다의 mAP, loss 값을 통해 비교하였으며, 학습 Phase에 따라 모델 고도화로 Test Set의 mAP 값이 성상별로 높아짐을 확인하였으며, 차집량 분석을 위해 Unet을 활용하여 Train, Test, Validation DataSet(8.5:1:0.5)을 구성하고 epoch별 IoU(intersection over Union), F1-score, loss 값을 비교하여 정성적, 정량적 평가 모두 Phase 3에서 가장 높은 성능을 확인하였다. 향후 하천 환경에서의 다양한 영양인자별 분석을 통해 주요 영향인자 도출 및 Hyper Parameter 최적화를 통한 모델 고도화로 인해 활용성이 높아질 것으로 판단된다.

• Smart Media Journal
• /
• v.11 no.7
• /
• pp.61-69
• /
• 2022

In the learned machine learning, the performance of machine learning degrades at the same time as drift occurs in terms of learning models and learning data over time. As a solution to this problem, I would like to propose the concept and evaluation method of ML drift to determine the re-learning period of machine learning. An XAI test and an XAI test of an apple image were performed according to strawberry and clarity. In the case of strawberries, the change in the XAI analysis of ML models according to the clarity value was insignificant, and in the case of XAI of apple image, apples normally classified objects and heat map areas, but in the case of apple flowers and buds, the results were insignificant compared to strawberries and apples. This is expected to be caused by the lack of learning images of apple flowers and buds, and more apple flowers and buds will be studied and tested in the future.

• Journal of Environmental Health Sciences
• /
• v.49 no.6
• /
• pp.295-304
• /
• 2023

Background: Adolescents are relatively more sensitive than adults to exposure to indoor pollutants. The indoor air quality in classrooms where students spend time together must therefore be managed at a safe level because it can affect the health of students. Objectives: In this study, three types of green-wall models were applied to classrooms where students spend a long time in a limited space, and the resulting effects on reducing PM were evaluated. Methods: In the middle school classrooms which were selected as the experimental subjects, IoT-based indoor air quality monitoring equipment was installed for real-time monitoring. Three types of plant models (passive, active, and active+light) were installed in each classroom to evaluate the effects on improving indoor air quality. Results: The concentration of PM in the classroom is influenced by outdoor air quality, but repeated increases and decreases in concentration were observed due to the influence of students' activities. There was a PM reduction effect by applying the green-wall model. There was a difference in PM reduction efficiency depending on the type of green-wall model, and the reduction efficiency of the active model was higher than the passive model. Conclusions: The active green-wall model can be used as an efficient method of improving indoor air quality. Additionally, more research is needed to increase the efficiency of improving indoor air quality by setting conditions that can stimulate the growth of each type of plant.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.19 no.1
• /
• pp.47-55
• /
• 2024

We spend a lot of time in indoor space, and the space has a huge impact on our lives. Interior design plays a significant role to make an indoor space attractive and functional. However, it should consider a lot of complex elements such as color, pattern, and material etc. With the increasing demand for interior design, there is a growing need for technologies that analyze these design elements accurately and efficiently. To address this need, this study suggests a deep learning-based design analysis system. The proposed system consists of a semantic segmentation model that classifies spatial components and an image classification model that classifies attributes such as color, pattern, and material from the segmented components. Semantic segmentation model was trained using a dataset of 30000 personal indoor interior images collected for research, and during inference, the model separate the input image pixel into 34 categories. And experiments were conducted with various backbones in order to obtain the optimal performance of the deep learning model for the collected interior dataset. Finally, the model achieved good performance of 89.05% and 0.5768 in terms of accuracy and mean intersection over union (mIoU). In classification part convolutional neural network (CNN) model which has recorded high performance in other image recognition tasks was used. To improve the performance of the classification model we suggests an approach that how to handle data that has data imbalance and vulnerable to light intensity. Using our methods, we achieve satisfactory results in classifying interior design component attributes. In this paper, we propose indoor space design analysis system that automatically analyzes and classifies the attributes of indoor images using a deep learning-based model. This analysis system, used as a core module in the A.I interior recommendation service, can help users pursuing self-interior design to complete their designs more easily and efficiently.

• Korean Journal of Remote Sensing
• /
• v.39 no.2
• /
• pp.193-205
• /
• 2023

A large amount of floating debris from land-based sources during heavy rainfall has negative social, economic, and environmental impacts, but there is a lack of monitoring systems for floating debris accumulation areas and amounts. With the recent development of artificial intelligence technology, there is a need to quickly and efficiently study large areas of water systems using drone imagery and deep learning-based object detection models. In this study, we acquired various images as well as drone images and trained with You Only Look Once (YOLO)v5s and the recently developed YOLO7 and YOLOv8s to compare the performance of each model to propose an efficient detection technique for land-based floating debris. The qualitative performance evaluation of each model showed that all three models are good at detecting floating debris under normal circumstances, but the YOLOv8s model missed or duplicated objects when the image was overexposed or the water surface was highly reflective of sunlight. The quantitative performance evaluation showed that YOLOv7 had the best performance with a mean Average Precision (intersection over union, IoU 0.5) of 0.940, which was better than YOLOv5s (0.922) and YOLOv8s (0.922). As a result of generating distortion in the color and high-frequency components to compare the performance of models according to data quality, the performance degradation of the YOLOv8s model was the most obvious, and the YOLOv7 model showed the lowest performance degradation. This study confirms that the YOLOv7 model is more robust than the YOLOv5s and YOLOv8s models in detecting land-based floating debris. The deep learning-based floating debris detection technique proposed in this study can identify the spatial distribution of floating debris by category, which can contribute to the planning of future cleanup work.

• The Korean Journal of Pharmacology
• /
• v.24 no.2
• /
• pp.203-216
• /
• 1988

To delineate the mechanisms of vasoconstriction and vasodilation in cerebral arteries the effects of some vasoconstrictors and calcium antagonists on the basilar artery (BA) and arterial circle of Willis (WC) were examined and also the role of endothelium in the action of these drugs was investigated in pigs, cats and rabbits. In pig cerebral arteries, dose-dependent contractile responses were elicited by KCI, histamine, 5-hydroxytryptamine (5-HT) and angiotensin, but norepinephrine (NE), phenylephrine (PE) and epinephrine (EP) elicited dose-dependent contractions only under pretreatment with propranolol 10-6 M. The magnitudes of maximal contractile effects of these drugs were different from each other, and 5-H~ was the largest and angiotensin the smallest. Some calcium antagonists dose-dependently inhibited KCI (35 mM)-induced contraction and the order of potency in inhibiting the contraction was nifedipine > > diltiazem > flunarizine > oxybutynin > isosorbide dinitrate (ISDN) > glyceryl trinitrate. 5-HT (10-6 M)-induced contraction was dosedependently inhibited by nifedipine but slightly inhibited by diltiazem and ISDN. In rings with intact endothelium, KCI (35 mM)-induced contraction was not affected by acetylcholine (ACh) but $PGF_{2{\alpha}}$ (lO-SM)-induced contraction was dose-dependently relaxed by ACh and adenosine. This endothelium-dependent relaxation was not affected by nifedipine (l0-6M)-pretreatment but markedly inhibited by methylene blue (50,uM)-pretreatment. In the porcine arterial rings without endothelium, ACh had no effect or even contracted the $PGF_{2{\alpha}}-induced$ contraction. However, the dosedependent relaxing effect of ACh appeared when the deendothelized porcine ring and rabbit thoracic aorta with intact endotheli urn were simultaneously suspended into a bath and this relaxing effect was also inhibited by methylene blue-pretreatment. In cat cerebral arteries, 5-HT and NE elicited dose-dependent contractile responses and ACh also produced dose-dependent contraction regardless of the existence of endothelium. ACh-induced contraction was most prominent. 5-HT (IO-SM)induced contraction was not relaxed but contracted additionally by ACh even in the intact endothelial ring. In rabbit cerebral arteries, 5-HT and NE elicited dose-dependent contractile responses and 5-HT-induced contraction was more prominent. In the intact endothelial preparations, 5-HT (lO-s M)-induced contraction was markedly relaxed by the addition of ACh( IO-SM) and this endothelium-dependent relaxing effect was inhibited by atropine (l0-7M)-pretreatment but notaffected by diltiazem (l0-6M)-pretreatment. These results suggest that ACh elicits endotheliumdependent relaxing effect mediated by muscarinic receptors in cerebral arteries of pig and rabbit, and that ACh acts as vasoconstrictor in cat cerebral artery.

• Journal of Aerospace System Engineering
• /
• v.14 no.1
• /
• pp.44-53
• /
• 2020

In conjunction with the Fourth Industrial Revolution, the unmanned aerial vehicle industry is being developed to a new paradigm by combining advanced technologies such as AI, Big Data and the IoT. Aeronautical developed countries such as the U.S. are focusing their efforts on the development of the safer unmanned aerial vehicles. The Korea Aerospace Research Institute, as part of the national R&D project in 2011, had succeeded in developing the first vertical takeoff and landing (VTOL) UAS, called Smart-UAV. However, although the development technology of the VTOL UAS is possessed, developing and operating of the VTOL UAS for commercial or military use are limited. The type certification procedure of the VTOL UAS developed by domestic technology is stipulated in the Korean Aviation Safety Act, but the Korean VTOL UAS airworthiness standards (KAS) hsve not been established. Thus, this study investigated the development trends of the VTOL UAS in Korea and abroad and national certification systems and procedures, and benchmarked the special conditions for the VTOL aircraft, announced by the EASA on July 2, 2019, to establish standards for type certificate of the VTOL UAS in Korea.

• The KIPS Transactions:PartC
• /
• v.9C no.6
• /
• pp.935-944
• /
• 2002

It is the Mobile Ad Hoc Networks that constituted with serveral mobile node that can communicate with other mobile nodes. Until now, there were no routing protocols considering such as Multimediadata, VOD (Video On Demand), which is required of lots of bandwidth in Mobile Ad Hoc Network, io we are in the need of QoS (Quality of Service)routing protocol to transmit the data packets faster and more accurate. Also, there are an unidirectional links due to asymmetric property of mobile terminals or current wireless environments on practical mobile ad hoc networks. However, at present, the existing mobile ad hoc routing protocols are implemented to support only bidirectional links. In this paper, we propose the Advanced Routing routing protocol in order to implement a new routing protocol, which is fit to mobile ad hoc networks containing unidirectional links and to support QoS service. For the performance evaluation, we use NS-2 simulator of U.C. Berkeley. We could get not only increased received data rate and decreased average route discovery time, but also network load decreases with compared Best effort service.

• Investigative Magnetic Resonance Imaging
• /
• v.7 no.1
• /
• pp.39-46
• /
• 2003

Purpose : To develop a theoretical model for magnetic relaxation behavior of the superparamagnetic nano-particle agent, which demonstrates multi-functionality such as liver- and lymp node-specificity. Based on the developed model, the computer simulation was performed to clarify the relationship between relaxation time and the applied magnetic field strength. Materials and Methods : The ultrasmall superparamagnetic iron oxide (USPIO) was encapsulated with biocompatiable polymer, to develop a relaxation model based on outsphere mechanism, which was resulting from diffusion and/or electron spin fluctuation. In addition, Brillouin function was introduced to describe the full magnetization by considering the fact that the low-field approximation, which was adapted in paramagnetic case, is no longer valid. The developed model describes therefore the T1 and T2 relaxation behavior of superparamagnetic iron oxide both in low-field and in high-field. Based on our model, the computer simulation was performed to test the relaxation behavior of superparamagnetic contrast agent over various magnetic fields using MathCad (MathCad, U.S.A.), a symbolic computation software. Results : For T1 and T2 magnetic relaxation characteristics of ultrasmall superparamagnetic iron oxide, the theoretical model showed that at low field (<1.0 Mhz), $\tau_{S1}(\tau_{S2}$, in case of T2), which is a correlation time in spectral density function, plays a major role. This suggests that realignment of nano-magnetic particles is most important at low magnetic field. On the other hand, at high field, $\tau$, which is another correlation time in spectral density function, plays a major role. Since $\tau$ is closely related to particle size, this suggests that the difference in R1 and R2 over particle sizes, at high field, is resulting not from the realignment of particles but from the particle size itself. Within normal body temperature region, the temperature dependence of T1 and T2 relaxation time showed that there is no change in T1 and T2 relaxation times at high field. Especially, T1 showed less temperature dependence compared to T2. Conclusion : We developed a theoretical model of r magnetic relaxation behavior of ultrasmall superparamagnetic iron oxide (USPIO), which was reported to show clinical multi-functionality by utilizing physical properties of nano-magnetic particle. In addition, based on the developed model, the computer simulation was performed to investigate the relationship between relaxation time of USPIO and the applied magnetic field strength.