• Journal of Internet Computing and Services
• /
• v.21 no.4
• /
• pp.109-116
• /
• 2020

With the advent of high-speed railways, railways are one of the most frequently used means of transportation at home and abroad. In addition, in terms of environment, carbon dioxide emissions are lower and energy efficiency is higher than other transportation. As the interest in railways increases, the issue related to railway safety is one of the important concerns. Among them, visual abnormalities occur when various obstacles such as animals and people suddenly appear in front of the railroad. To prevent these accidents, detecting rail tracks is one of the areas that must basically be detected. Images can be collected through cameras installed on railways, and the method of detecting railway rails has a traditional method and a method using deep learning algorithm. The traditional method is difficult to detect accurately due to the various noise around the rail, and using the deep learning algorithm, it can detect accurately, and it combines the two algorithms to detect the exact rail. The proposed algorithm determines the accuracy of railway rail detection based on the data collected.

• Korean Journal of Remote Sensing
• /
• v.38 no.5_2
• /
• pp.793-810
• /
• 2022

Synthetic Aperture Radar (SAR) is considered to be suitable for near real-time inundation monitoring. The distinctly different intensity between water and land makes it adequate for waterbody detection, but the intrinsic speckle noise and variable intensity of SAR images decrease the accuracy of waterbody detection. In this study, we suggest two modules, named 'morphology module' and 'edge-enhanced module', which are the combinations of pooling layers and convolutional layers, improving the accuracy of waterbody detection. The morphology module is composed of min-pooling layers and max-pooling layers, which shows the effect of morphological transformation. The edge-enhanced module is composed of convolution layers, which has the fixed weights of the traditional edge detection algorithm. After comparing the accuracy of various versions of each module for U-Net, we found that the optimal combination is the case that the morphology module of min-pooling and successive layers of min-pooling and max-pooling, and the edge-enhanced module of Scharr filter were the inputs of conv9. This morphologic and edge-enhanced U-Net improved the F1-score by 9.81% than the original U-Net. Qualitative inspection showed that our model has capability of detecting small-sized waterbody and detailed edge of water, which are the distinct advancement of the model presented in this research, compared to the original U-Net.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.513-515
• /
• 2022

Manufacturing data collected using IoT devices in a smart factory environment is generally reliable except for noises caused by external factors. However, unlike manufacturing data that is collected mechanically, process data manually recorded by field-workers can cause problems such as the misspelled entries or the missing entries. Therefore, process data must be validated before being used as training data for artificial intelligence models. In this paper, based on the fact that which is a linear relationship between the power consumption of the MCT machine and the production of the product recorded by the field-workers, we detect anomalous records of the workers using robust linear regression.

• Journal of the Korean earth science society
• /
• v.42 no.6
• /
• pp.623-631
• /
• 2021

The increased transport of marine hazardous and noxious substances (HNS) has resulted in frequent HNS spill accidents domestically and internationally. There are about 6,000 species of HNS internationally, and most of them have toxic properties. When an accidental HNS spill occurs, it can destroys the marine ecosystem and can damage life and property due to explosion and fire. Constructing a spectral library of HNS according to wavelength and developing a detection algorithm would help prepare for accidents. In this study, a ground HNS spill experiment was conducted in France. The toluene spectrum was determined through hyperspectral sensor measurements. HNS present in the hyperspectral images were detected by applying the spectral mixture algorithm. Preprocessing principal component analysis (PCA) removed noise and performed dimensional compression. The endmember spectra of toluene and seawater were extracted through the N-FINDR technique. By calculating the abundance fraction of toluene and seawater based on the spectrum, the detection accuracy of HNS in all pixels was presented as a probability. The probability was compared with radiance images at a wavelength of 418.15 nm to select abundance fractions with maximum detection accuracy. The accuracy exceeded 99% at a ratio of approximately 42%. Response to marine spills of HNS are presently impeded by the restricted access to the site because of high risk of exposure to toxic compounds. The present experimental and detection results could help estimate the area of contamination with HNS based on hyperspectral remote sensing.

• Korean Journal of Remote Sensing
• /
• v.29 no.4
• /
• pp.351-360
• /
• 2013

Harmonic analysis enables to characterize patterns of variation in MODIS NDVI time series data and track changes in ground vegetation cover. In harmonic analysis, a periodic phenomenon of time series data is decomposed into the sum of a series of sinusoidal waves and an additive term. Each wave is defined by an amplitude and a phase angle and accounts for the portion of variance of complex curve. In this study, harmonic analysis was explored to tract ground vegetation variation through time for land-cover vegetation change detection. The process also enables to reconstruct observed time series data including various noise components. Harmonic model was tested with simulation data to validate its performance. Then, the suggested change detection method was applied to MODIS NDVI time series data over the study period (2006-2012) for a selected test area located in the northern plateau of Korean peninsula. The results show that the proposed approach is potentially an effective way to understand the pattern of NDVI variation and detect the change for long-term monitoring of land cover.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.48 no.5
• /
• pp.45-51
• /
• 2011

In this paper, we propose an unsupervised learning method for modeling motion trajectory patterns effectively. In our approach, observations of an object on a trajectory are treated as words in a document for latent dirichlet allocation algorithm which is used for clustering words on the topic in natural language process. This allows clustering topics (e.g. go straight, turn left, turn right) effectively in complex scenes, such as crossroads. After this procedure, we learn patterns of word sequences in each cluster using Baum-Welch algorithm used to find the unknown parameters in a hidden markov model. Evaluation of abnormality can be done using forward algorithm by comparing learned sequence and input sequence. Results of experiments show that modeling of semantic region is robust against noise in various scene.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2015.05a
• /
• pp.275-275
• /
• 2015

최근 기후변화에 따른 극한 강우사상의 증가로 인하여 농업용 저수지의 재해 위험도가 증가하고 있는 추세이며, 사고가 발생할 때 마다 파손/붕괴된 시설물을 보수하는 대응형 유지관리체계에서 벗어나 기반시설의 성능과 생애주기 등을 고려하여 재해 발생을 사전에 예보 및 경보를 알릴 수 있는 예방적 관리체계로의 전환이 필요하다. 한국농어촌공사는 전국 1,500개 저수지에서 10분 단위 수위자료를 측정하고 있으며, 이를 분석하여 재해예방에 활용할 수 있는 기반이 조성되어 있으나 이에 대한 관리가 이루어지지 않고 있고 수집된 자료를 활용하여 재해 징후를 분석할 수 있는 재해 예방적 분석기술이 마련되어 있지 않은 실정이다. 본 연구에서는 농업용 저수지 수위자료를 이용한 저수지 이상거동을 판별하기 위하여 전국 34개 한국농어촌공사 관할 저수의 시계열 수위자료의 특성(Feature)을 분석하고자 한다. 시계열 자료의 시계열 특성을 분석하기 위하여 한국농어촌공사 관할의 전국 34개 저수지를 선정하여 분석을 실시하였다. 대상저수지는 지역별, 저수용량, 안정등급, 붕괴발생, 1개 지사관할 저수지로 각각 구분하여 선정하였으며, 각 저수지의 수위 측정기간(최소 5개년)에 대한 자료를 수집하였다. 농업용 저수지의 시계열 수위 자료의 특성을 분석하기 위하여 자료의 전처리를 수행하였다. 자료의 전처리는 시계열 수위자료의 잡음 특성, 기상자료 관련 변동특성 등 분류(Classification)에 영향을 미치는 노이즈 요소를 제거하는 과정이다. 전처리과정을 거친 자료는 특징(Feature) 추출 과정을 거치게 되고, 추출된 특징의 적합성에 따라 분류 알고리듬 성능에 많은 영향을 미친다. 따라서 시계열 자료의 특성을 파악하고 특징을 추출하는 것은 이상치 탐지에 있어 매우 중요한 과정이다. 본 연구에서는 시계열 자료 특징 추출 방법으로 물리적인 한계치, 확률적인 문턱값(Threshold), 시계열 패턴, 주변 저수지와의 시계열 상관분석 등을 적용하였으며, 이를 데이터베이스로 구축하여 이후 분류알고리듬 학습에 적용하여 정상치와 이상치를 판별하는데 이용될 수 있도록 하였다. 따라서 본 연구에서 제시되는 농업용 저수지의 시계열 특성은 다양한 분류알고리듬에 적용할 수 있으며, 이를 통하여 저수지 이상거동 판별을 위한 최적을 분류알고리듬의 선택에 도움이 될 것이다.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.6
• /
• pp.907-917
• /
• 2023

Recently, a self-driving car have applied deep learning technology to advanced driver assistance system can provide convenience to drivers, but it is shown deep that learning technology is vulnerable to adversarial evasion attacks. In this paper, we performed five adversarial evasion attacks, including MI-FGSM(Momentum Iterative-Fast Gradient Sign Method), targeting the object detection algorithm YOLOv5 (You Only Look Once), and measured the object detection performance in terms of mAP(mean Average Precision). In particular, we present a method applying morphology operations for YOLO to detect objects normally by removing noise and extracting boundary. As a result of analyzing its performance through experiments, when an adversarial attack was performed, YOLO's mAP dropped by at least 7.9%. The YOLO applied our proposed method can detect objects up to 87.3% of mAP performance.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.10
• /
• pp.67-76
• /
• 2023

Although deep learning models are making innovative achievements in the field of computer vision, the problem of vulnerability to adversarial examples continues to be raised. Adversarial examples are attack methods that inject fine noise into images to induce misclassification, which can pose a serious threat to the application of deep learning models in the real world. In this paper, we propose a model that detects adversarial examples using differences in predictive values between edge-learned classification models and underlying classification models. The simple process of extracting the edges of the objects and reflecting them in learning can increase the robustness of the classification model, and economical and efficient detection is possible by detecting adversarial examples through differences in predictions between models. In our experiments, the general model showed accuracy of {49.9%, 29.84%, 18.46%, 4.95%, 3.36%} for adversarial examples (eps={0.02, 0.05, 0.1, 0.2, 0.3}), whereas the Canny edge model showed accuracy of {82.58%, 65.96%, 46.71%, 24.94%, 13.41%} and other edge models showed a similar level of accuracy also, indicating that the edge model was more robust against adversarial examples. In addition, adversarial example detection using differences in predictions between models revealed detection rates of {85.47%, 84.64%, 91.44%, 95.47%, and 87.61%} for each epsilon-specific adversarial example. It is expected that this study will contribute to improving the reliability of deep learning models in related research and application industries such as medical, autonomous driving, security, and national defense.

• Journal of Internet Computing and Services
• /
• v.25 no.1
• /
• pp.109-121
• /
• 2024

In various manufacturing processes such as textiles and automobiles, when equipment breaks down or stops, the machines do not work, which leads to time and financial losses for the company. Therefore, it is important to detect equipment abnormalities in advance so that equipment failures can be predicted and repaired before they occur. Most equipment failures are caused by bearing failures, which are essential parts of equipment, and detection bearing anomaly is the essence of PHM(Prognostics and Health Management) research. In this paper, we propose a preprocessing algorithm called SWT-SVD, which analyzes vibration signals from bearings and apply it to an anomaly transformer, one of the time series anomaly detection model networks, to implement bearing anomaly detection model. Vibration signals from the bearing manufacturing process contain noise due to the real-time generation of sensor values. To reduce noise in vibration signals, we use the Stationary Wavelet Transform to extract frequency components and perform preprocessing to extract meaningful features through the Singular Value Decomposition algorithm. For experimental validation of the proposed SWT-SVD preprocessing method in the bearing anomaly detection model, we utilize the PHM-2012-Challenge dataset provided by the IEEE PHM Conference. The experimental results demonstrate significant performance with an accuracy of 0.98 and an F1-Score of 0.97. Additionally, to substantiate performance improvement, we conduct a comparative analysis with previous studies, confirming that the proposed preprocessing method outperforms previous preprocessing methods in terms of performance.