• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.1
• /
• pp.53-62
• /
• 2023

Due to the recent increase in new and renewable energy, gas turbine generators start and stop every day to supply high-quality power, and accordingly, the life span of high-temperature parts is shortened and the failure of combustion chamber temperature sensors increases. Therefore, in this study, we proposed a fuzzy logic-based failure diagnosis algorithm that can accurately diagnose and systematically detect the failure of the sensor when the dual temperature sensor used for gas turbine control fails, and to confirm the usefulness of the proposed algorithm We tried to confirm the usefulness of the proposed algorithm by performing various simulations under the matlab/simulink environment.

• International Journal of Computer Science & Network Security
• /
• v.23 no.8
• /
• pp.49-62
• /
• 2023

There has been a rapid increase in the creation and alteration of new malware samples which is a huge financial risk for many organizations. There is a huge demand for improvement in classification and detection mechanisms available today, as some of the old strategies like classification using mac learning algorithms were proved to be useful but cannot perform well in the scalable auto feature extraction scenario. To overcome this there must be a mechanism to automatically analyze malware based on the automatic feature extraction process. For this purpose, the dynamic analysis of real malware executable files has been done to extract useful features like API call sequence and opcode sequence. The use of different hashing techniques has been analyzed to further generate images and convert them into image representable form which will allow us to use more advanced classification approaches to classify huge amounts of images using deep learning approaches. The use of deep learning algorithms like convolutional neural networks enables the classification of malware by converting it into images. These images when fed into the CNN after being converted into the grayscale image will perform comparatively well in case of dynamic changes in malware code as image samples will be changed by few pixels when classified based on a greyscale image. In this work, we used VGG-16 architecture of CNN for experimentation.

• Applied Microscopy
• /
• v.52
• /
• pp.1.1-1.8
• /
• 2022

As semiconductor device architecture develops, from planar field-effect transistors (FET) to FinFET and gate-all-around (GAA), there is an increased need to measure 3D structure sidewalls precisely. Here, we present a 3-Dimensional Atomic Force Microscope (3D-AFM), a powerful 3D metrology tool to measure the sidewall roughness (SWR) of vertical and undercut structures. First, we measured three different dies repeatedly to calculate reproducibility in die level. Reproducible results were derived with a relative standard deviation under 2%. Second, we measured 13 different dies, including the center and edge of the wafer, to analyze SWR distribution in wafer level and reliable results were measured. All analysis was performed using a novel algorithm, including auto fattening, sidewall detection, and SWR calculation. In addition, SWR automatic analysis software was implemented to reduce analysis time and to provide standard analysis. The results suggest that our 3D-AFM, based on the tilted Z scanner, will enable an advanced methodology for automated 3D measurement and analysis.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.134-134
• /
• 2021

Precipitation is a crucial component of water cycle and play a key role in hydrological processes. Traditionally, gauge-based precipitation is the main method to achieve high accuracy of rainfall estimation, but its distribution is sparsely in mountainous areas. Recently, satellite-based precipitation products (SPPs) provide grid-based precipitation with spatio-temporal variability, but SPPs contain a lot of uncertainty in estimated precipitation, and the spatial resolution quite coarse. To overcome these limitations, this study aims to generate new grid-based daily precipitation using Automatic weather system (AWS) in Korea and multiple SPPs(i.e. CHIRPSv2, CMORPH, GSMaP, TRMMv7) during the period of 2003-2017. And this study used a machine learning based Random Forest (RF) model for generating new merging precipitation. In addition, several statistical linear merging methods are used to compare with the results of the RF model. In order to investigate the efficiency of RF, observed data from 64 observed Automated Synoptic Observation System (ASOS) were collected to evaluate the accuracy of the products through Kling-Gupta efficiency (KGE), probability of detection (POD), false alarm rate (FAR), and critical success index (CSI). As a result, the new precipitation generated through the random forest model showed higher accuracy than each satellite rainfall product and spatio-temporal variability was better reflected than other statistical merging methods. Therefore, a random forest-based ensemble satellite precipitation product can be efficiently used for hydrological simulations in ungauged basins such as the Mekong River.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.1249-1252
• /
• 2022

In this paper, we propose a conceptual framework that identifies video frames in motion containing the movement of people and vehicles in traffic videos. The automatic selection of video frames in motion is an important topic in security and surveillance video because the number of videos to be monitored simultaneously is simply too large due to limited human resources. The conventional method to identify the areas in motion is to compute the differences over consecutive video frames, which has been costly because of its high computational complexity. In this paper, we reduced the overall complexity by examining only the keyframes (or I-frames). The basic assumption is that the time period between I-frames is rather shorter (e.g., 1/10 ~ 3 secs) than the usual length of objects in motion in video (i.e., pedestrian walking, automobile passing, etc.). The proposed method estimates the possibility of videos containing motion between I-frames by evaluating the difference of consecutive I-frames with the long-time statistics of the previously decoded I-frames of the same video. The experimental results showed that the proposed method showed more than 80% accuracy in short surveillance videos obtained from different locations while keeping the computational complexity as low as 20 % compared to the HM decoder.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.493-494
• /
• 2023

본 논문은 벚꽃나무 영상 데이터를 활용하여 벚꽃의 상태(개화, 만개, 낙화)를 실시간으로 분류하는 연구를 소개한다. 이 연구의 목적은, 실시간으로 취득되는 벚꽃나무의 영상 데이터를 사전에 학습된 CNN 기반 이미지 분류 모델을 통해 벚꽃의 상태에 따라 분류하는 것이다. 약 1,000장의 벚꽃나무 이미지를 활용하여 CNN 모델을 학습시키고, 모델이 새로운 이미지에 대해 얼마나 정확하게 벚꽃의 상태를 분류하는지를 평가하였다. 학습데이터는 훈련 데이터와 검증 데이터로 나누었으며, 개화, 만개, 낙화 등의 상태별로 폴더를 구분하여 관리하였다. 또한, ImageNet 데이터셋에서 사전 학습된 ResNet50 가중치를 사용하는 전이학습 방법을 적용하여 학습 과정을 더 효율적으로 수행하고, 모델의 성능을 향상시켰다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.739-740
• /
• 2023

코로나19 의 여파로 생활 폐기물은 급속도로 늘어나는 반면 재활용 사업장의 여건은 개선되지 않고 있어 재활용 산업의 인력난 해결의 필요성이 떠오르고 있다. 이를 위해 본 논문에서는 딥러닝 모델을 활용하여 재활용 폐기물을 분류하는 방법을 제시한다. 딥러닝 모델은 최신 객체 탐지 모델인 YOLOv5를 사용하고, 객체 탐지 성능을 향상시키기 위해 실제 환경에서 수집된 학습용 데이터를 직접 라벨링하여 사용한다. 실험 결과 종류별 평균 0.69의 mAP50 스코어를 기록하였으며 이를 통해 딥러닝 모델을 활용하여 재활용 폐기물을 효율적으로 분류하는 것이 가능함을 확인하였다.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2022.11a
• /
• pp.186-186
• /
• 2022

과거에는 선박을 운용하기 위해서 많은 인원이 필요하였으나 최근 들어 선박 운용에 필요한 인원이 줄어들고 있으며, 더 나아가 자율적으로 운항하는 선박을 만들기 위한 연구가 활발히 수행되고 있다. 자율 운항 선박을 구성하는 여러 요소 중 인간의 시각을 대체하기 위한 자율 인지 시스템은 가장 선행되어야 하는 연구 분야 중 하나이다. RADAR (RAdio Detection And Ranging) 및 AIS (Automatic Identification System) 등의 전통적인 인지 센서를 활용한 연구가 진행 중이지만 사각지대나 탐지 주기 등의 한계가 있다. 따라서 본 연구에서는 다중 카메라 (광학, 열상, 파노라마)를 이용하여 전통적인 인지 센서의 한계를 보완하는 새로운 인지 시스템을 고안하였으며, 이를 기반으로 해상 장애물을 추적하여 동적 운동 정보를 얻었다. 먼저 실해역에서 수집한 이미지를 바탕으로 해상 장애물 탐지를 위한 데이터를 구성하고, 딥 러닝 기반의 탐지 모델을 학습시켰다. 탐지 모델을 이용하여 탐지한 결과는 직접 설계한 칼만 필터 기반의 적응형 추적 필터를 통과시켜 해상 장애물의운동 정보 (궤적, 속력, 방향)를 계산하는데 활용되었다. 또한 본 연구는 카메라를 센서로 활용했을 때의 한계를 보완하기 위하여 동 시간대에 다중 카메라에서 추적한 각각의 정보를 융합하였다. 그 결과 단일 카메라를 활용하는 경우, RADAR의 오차 범위 이내에 추적 결과가 수렴하는 양상을 보였으며, 다중 카메라를 활용하는 경우에는 단일 카메라보다 정확한 추적이 가능함을 확인하였다.

• Smart Structures and Systems
• /
• v.33 no.1
• /
• pp.41-53
• /
• 2024

Grouting sleeves are an essential connecting component of prefabricated components, and the quality of grouting has a significant influence on structural integrity and seismic performance. The embedded grouting sleeve (EGS)'s grouting defects are highly undetectable and random, and no effective monitoring method exists. This paper proposes an ultrasonic guided wave method and provides a set of guidelines for selecting the optimal frequency and suitable period for the EGS. The optimal frequency was determined by considering the group velocity, wave structure, and wave attenuation of the selected mode. Guided waves are prone to multi-modality, modal conversion, energy leakage, and dispersion in the EGS, which is a multi-layer structure. Therefore, a time-reversal (TR)-based multi-mode focusing and dispersion automatic compensation technology is introduced to eliminate the multi-mode phase difference in the EGS. First, the influence of defects on guided waves is analyzed according to the TR coefficient. Second, two major types of damage indicators, namely, the time domain and the wavelet packet energy, are constructed according to the influence method. The constructed wavelet packet energy indicator is more sensitive to the changes of defecting than the conventional time-domain similarity indicator. Both numerical and experimental results show that the proposed method is feasible and beneficial for the detection and quantitative estimation of the grouting defects of the EGS.

• Advances in environmental research
• /
• v.12 no.2
• /
• pp.113-122
• /
• 2023

Biometric authentication has become an essential part of modern-day security systems, especially in financial institutions like banks. A face recognition-based ATM is a biometric authentication system, that uses facial recognition technology to verify the identity of bank account holders during ATM transactions. This technology offers a secure and convenient alternative to traditional ATM transactions that rely on PIN numbers for verification. The proposed system captures users' pictures and compares it with the stored image in the bank's database to authenticate the transaction. The technology also offers additional benefits such as reducing the risk of fraud and theft, as well as speeding up the transaction process. However, privacy and data security concerns remain, and it is important for the banking sector to instrument solid security actions to protect customers' personal information. The proposed system consists of two stages: the first stage captures the user's facial image using a camera and performs pre-processing, including face detection and alignment. In the second stage, machine learning algorithms compare the pre-processed image with the stored image in the database. The results demonstrate the feasibility and effectiveness of using face recognition for ATM authentication, which can enhance the security of ATMs and reduce the risk of fraud.