• 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 the Korean Society of Marine Environment & Safety
• /
• v.29 no.6
• /
• pp.536-542
• /
• 2023

Satellites represent cutting-edge technology, of ering significant advantages in spatial and temporal observations. National agencies worldwide harness satellite data to respond to marine accidents and analyze ocean fluctuations effectively. However, challenges arise with high-resolution satellite-based sea surface temperature data (Operational Sea Surface Temperature and Sea Ice Analysis, OSTIA), where gaps or empty areas may occur due to satellite instrumentation, geographical errors, and cloud cover. These issues can take several hours to rectify. This study addressed the issue of missing OSTIA data by employing LaMa, the latest deep learning-based algorithm. We evaluated its performance by comparing it to three existing image processing techniques. The results of this evaluation, using the coefficient of determination (R²) and mean absolute error (MAE) values, demonstrated the superior performance of the LaMa algorithm. It consistently achieved R² values of 0.9 or higher and kept MAE values under 0.5 ℃ or less. This outperformed the traditional methods, including bilinear interpolation, bicubic interpolation, and DeepFill v1 techniques. We plan to evaluate the feasibility of integrating the LaMa technique into an operational satellite data provision system.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.8
• /
• pp.958-964
• /
• 2016

In this paper, we propose an i-vector based language recognition system to identify the spoken language of the speaker, which uses multiple discriminative deep neural network (DNN) models analogous to the multi-class support vector machine (SVM) classification system. The proposed model was trained and tested using the i-vectors included in the NIST 2015 i-vector Machine Learning Challenge database, and shown to outperform the conventional language recognition methods such as cosine distance, SVM and softmax NN classifier in open-set experiments.

• Journal of Korean Association for Spatial Structures
• /
• v.20 no.2
• /
• pp.51-58
• /
• 2020

Recently, deep learning that is the most popular and effective class of machine learning algorithms is widely applied to various industrial areas. A number of research on various topics about structural engineering was performed by using artificial neural networks, such as structural design optimization, vibration control and system identification etc. When nonlinear semi-active structural control devices are applied to building structure, a lot of computational effort is required to predict dynamic structural responses of finite element method (FEM) model for development of control algorithm. To solve this problem, an artificial neural network model was developed in this study. Among various deep learning algorithms, a recurrent neural network (RNN) was used to make the time history response prediction model. An RNN can retain state from one iteration to the next by using its own output as input for the next step. An eleven-story building structure with semi-active tuned mass damper (TMD) was used as an example structure. The semi-active TMD was composed of magnetorheological damper. Five historical earthquakes and five artificial ground motions were used as ground excitations for training of an RNN model. Another artificial ground motion that was not used for training was used for verification of the developed RNN model. Parametric studies on various hyper-parameters including number of hidden layers, sequence length, number of LSTM cells, etc. After appropriate training iteration of the RNN model with proper hyper-parameters, the RNN model for prediction of seismic responses of the building structure with semi-active TMD was developed. The developed RNN model can effectively provide very accurate seismic responses compared to the FEM model.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.1
• /
• pp.35-40
• /
• 2022

In this paper, we propose a method of constructing clinical knowledge with clinical concept relations and predicting diseases based on a deep learning model to support clinical decision-making. Clinical terms in UMLS(Unified Medical Language System) and cancer-related medical knowledge are classified into five categories. Medical related documents in Wikipedia are extracted using the classified clinical terms. Clinical concept relations are established by matching the extracted medical related documents with the extracted clinical terms. After deep learning using clinical knowledge, a disease is predicted based on medical terms expressed in a query. Thereafter, medical terms related to the predicted disease are selected as an extended query for clinical document retrieval. To validate our method, we have experimented on TREC Clinical Decision Support (CDS) and TREC Precision Medicine (PM) test collections.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.12
• /
• pp.1853-1858
• /
• 2021

O-rings fill the gaps between mechanical parts. Until now, the sorting of defective products has been performed visually and manually, so classification errors often occur. Therefore, a camera-based defect classification system without human intervention is required. However, a binarization process is required to separate the required region from the background in the camera input image. In this paper, an adaptive binarization technique that considers the surrounding pixel values is applied to solve the problem that single-threshold binarization is difficult to apply due to factors such as changes in ambient lighting or reflections. In addition, the convex hull technique is also applied to compensate for the missing pixel part. And the learning model to be applied to the separated region applies the residual error-based deep learning neural network model, which is advantageous when the defective characteristic is non-linear. It is suggested that the proposed system through experiments can be applied to the automation of O-ring defect detection.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.4
• /
• pp.54-62
• /
• 2018

As a part of realization of artificial intelligence signal(AI Signal), this study proposed an actuated signal algorithm based on vehicle queue length that estimates in real time by deep learning. In order to implement the algorithm, we built an API(COM Interface) to control the micro traffic simulator Vissim in the tensorflow that implements the deep learning model. In Vissim, when the link travel time and the traffic volume collected by signal cycle are transferred to the tensorflow, the vehicle queue length is estimated by the deep learning model. The signal time is calculated based on the vehicle queue length, and the simulation is performed by adjusting the signaling inside Vissim. The algorithm developed in this study is analyzed that the vehicle delay is reduced by about 5% compared to the current TOD mode. It is applied to only one intersection in the network and its effect is limited. Future study is proposed to expand the space such as corridor control or network control using this algorithm.

• Korean Journal of Remote Sensing
• /
• v.38 no.5_3
• /
• pp.901-912
• /
• 2022

The frequency of disasters is increasing due to global climate change, and unusual heavy rains and rainy seasons are occurring in Korea. Periodic monitoring and rapid detection are important because these weather conditions can lead to drought and flooding, causing secondary damage. Although research using optical images is continuously being conducted to determine the waterbody, there is a limitation in that it is difficult to detect due to the influence of clouds in order to detect floods that accompany heavy rain. Therefore, there is a need for research using synthetic aperture radar (SAR) that can be observed regardless of day or night in all weather. In this study, using Sentinel-1 SAR images that can be collected in near-real time as open data, the UNet model among deep learning algorithms that have recently been used in various fields was applied. In previous studies, waterbody detection studies using SAR images and deep learning algorithms are being conducted, but only a small number of studies have been conducted in Korea. In this study, to determine the applicability of deep learning of SAR images, UNet and the existing algorithm thresholding method were compared, and five indices and Sentinel-2 normalized difference water index (NDWI) were evaluated. As a result of evaluating the accuracy with intersect of union (IoU), it was confirmed that UNet has high accuracy with 0.894 for UNet and 0.699 for threshold method. Through this study, the applicability of deep learning-based SAR images was confirmed, and if high-resolution SAR images and deep learning algorithms are applied, it is expected that periodic and accurate waterbody change detection will be possible in Korea.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.2
• /
• pp.517-524
• /
• 2024

This study aims to train and implement a deep learning model for the fusion of website creation and artificial intelligence, in the era known as the AI revolution following the launch of the ChatGPT service. The deep learning model was trained using 3,000 collected web page images, processed based on a system of component and layout classification. This process was divided into three stages. First, prior research on AI models was reviewed to select the most appropriate algorithm for the model we intended to implement. Second, suitable web page and paragraph images were collected, categorized, and processed. Third, the deep learning model was trained, and a serving interface was integrated to verify the actual outcomes of the model. This implemented model will be used to detect multiple paragraphs on a web page, analyzing the number of lines, elements, and features in each paragraph, and deriving meaningful data based on the classification system. This process is expected to evolve, enabling more precise analysis of web pages. Furthermore, it is anticipated that the development of precise analysis techniques will lay the groundwork for research into AI's capability to automatically generate perfect web pages.

• Smart Media Journal
• /
• v.9 no.3
• /
• pp.25-30
• /
• 2020

The purpose of this paper is to implement a deep learning-based real-time video analysis algorithm that monitors safety of workers in industrial facilities. The worker's clothes were divided into six classes according to whether workers are wearing a helmet, safety vest, and safety belt, and a total of 5,307 images were used as learning data. The experiment was performed by comparing the mAP when weight was applied according to the number of learning iterations for 645 images, using YOLO v4. It was confirmed that the mAP was the highest with 60.13% when the number of learning iterations was 6,000, and the AP with the most test sets was the highest. In the future, we plan to improve accuracy and speed by optimizing datasets and object detection model.