• Journal of the Korean Society of Marine Environment & Safety
• /
• v.26 no.7
• /
• pp.759-766
• /
• 2020

In recent years, the maritime traffic environment has been changing in various ways, and the traffic volume has been increasing constantly. Accordingly, the requirements for maritime traffic analysis have become diversified. To this end, traffic characteristics must first be analyzed using vessel trajectory data. However, as the conventional method is mostly manual, it requires a considerable amount of time and effort, and errors may occur during data processing. In addition, ensuring the reliability of the analysis results is difficult, because this method considers the subjective opinion of analysts. Therefore, in this paper, we propose an automated method of traffic network generation for maritime traffic analysis. In the experiment, spatiotemporal features are analyzed using data collected at Mokpo Harbor over six months. The proposed method can automatically generate a traffic network reflecting the traffic characteristics of the experimental area. In addition, it can be applied to a large amount of trajectory data. Finally, as the spatiotemporal characteristics can be analyzed using the traffic network, the proposed method is expected to be used in various maritime traffic analyses.

• Korean Journal of Remote Sensing
• /
• v.38 no.1
• /
• pp.111-126
• /
• 2022

Land use and land cover (LULC) mapping is an important factor in geospatial analysis. Although highly precise ground-based LULC monitoring is possible, it is time consuming and costly. Conversely, because the synthetic aperture radar (SAR) sensor is an all-weather sensor with high resolution, it could replace field-based LULC monitoring systems with low cost and less time requirement. Thus, LULC is one of the major areas in SAR applications. We developed a LULC model using only KOMPSAT-5 single co-polarized data and digital elevation model (DEM) data. Twelve HH-polarized images and 18 VV-polarized images were collected, and two HH-polarized images and four VV-polarized images were selected for the model testing. To train the LULC model, we applied the conditional generative adversarial network (cGAN) method. We used U-Net combined with the residual unit (ResUNet) model to generate the cGAN method. When analyzing the training history at 1732 epochs, the ResUNet model showed a maximum overall accuracy (OA) of 93.89 and a Kappa coefficient of 0.91. The model exhibited high performance in the test datasets with an OA greater than 90. The model accurately distinguished water body areas and showed lower accuracy in wetlands than in the other LULC types. The effect of the DEM on the accuracy of LULC was analyzed. When assessing the accuracy with respect to the incidence angle, owing to the radar shadow caused by the side-looking system of the SAR sensor, the OA tended to decrease as the incidence angle increased. This study is the first to use only KOMPSAT-5 single co-polarized data and deep learning methods to demonstrate the possibility of high-performance LULC monitoring. This study contributes to Earth surface monitoring and the development of deep learning approaches using the KOMPSAT-5 data.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.25 no.2
• /
• pp.71-81
• /
• 2021

This paper presents a real-time, false-pick filter based on deep learning to reduce false alarms of an onsite Earthquake Early Warning (EEW) system. Most onsite EEW systems use P-wave to predict S-wave. Therefore, it is essential to properly distinguish P-waves from noises or other seismic phases to avoid false alarms. To reduce false-picks causing false alarms, this study made the EEWNet Part 1 'False-Pick Filter' model based on Convolutional Neural Network (CNN). Specifically, it modified the Pick_FP (Lomax et al.) to generate input data such as the amplitude, velocity, and displacement of three components from 2 seconds ahead and 2 seconds after the P-wave arrival following one-second time steps. This model extracts log-mel power spectrum features from this input data, then classifies P-waves and others using these features. The dataset consisted of 3,189,583 samples: 81,394 samples from event data (727 events in the Korean Peninsula, 103 teleseismic events, and 1,734 events in Taiwan) and 3,108,189 samples from continuous data (recorded by seismic stations in South Korea for 27 months from 2018 to 2020). This model was trained with 1,826,357 samples through balancing, then tested on continuous data samples of the year 2019, filtering more than 99% of strong false-picks that could trigger false alarms. This model was developed as a module for USGS Earthworm and is written in C language to operate with minimal computing resources.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.9
• /
• pp.167-176
• /
• 2023

In this paper, we designed and implemented a graph assessment module that can evaluate graphs in an programming assessment system based on text and numbers. The assessment method of the graph assessment module is self-evaluation that outputs two graphs generated by codes submitted by learners and by answers, automatic-evaluation that converts each graph image into an array, and gives feedback if it is wrong. The data used to generate the graph can be inputted directly or used from external data, and the method of generatng graph that can be evaluated is MATLAB style in matplotlib, and the graph shape that can be evaluated is presented in mathematics and curriculum. Through expert review, it was confirmed that the content elements of the assessment module, the possibility of learning, and the validity of the learner's needs were met. The graph assessment module developed in this study has expanded the evaluation area of the programming automatic asssessment system and is expected to help students learn data visualization.

• Tunnel and Underground Space
• /
• v.33 no.6
• /
• pp.495-507
• /
• 2023

The roughness of discontinuity significantly influences the mechanical characteristics of rock masses and extensively affects thermal and hydraulic behaviors. In this study, we utilized photogrammetry to generate 3D point cloud data for discontinuity and applied this data to characterize the roughness of discontinuity. The discontinuity profiles, reconstructed from the 3D point cloud data, were compared with those manually measured using a profile gauge. This comparison served to validate the accuracy and reliability of the acquired point cloud data in replicating the actual configurations of rock surfaces. Subsequent to this validation, influence of the number of profiles for representative JRC assessment was further investigated followed by suggestion of roughness anisotropy evaluation method with application of it to actual rock discontinuity surfaces.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.6
• /
• pp.113-120
• /
• 2024

With the advancement of information and communication technology, we can easily generate various forms of data in our daily lives. To efficiently manage such a large amount of data, systematic classification into categories is essential. For effective search and navigation, data is organized into a tree-like hierarchical structure known as a category tree, which is commonly seen in news websites and Wikipedia. As a result, various techniques have been proposed to classify large volumes of documents into the terminal nodes of category trees. However, document classification methods using category trees face a problem: as the height of the tree increases, the number of terminal nodes multiplies exponentially, which increases the probability of misclassification and ultimately leads to a reduction in classification accuracy. Therefore, in this paper, we propose a new node expansion-based classification algorithm that satisfies the classification accuracy required by the application, while enabling detailed categorization. The proposed method uses a greedy approach to prioritize the expansion of nodes with high classification accuracy, thereby maximizing the overall classification accuracy of the category tree. Experimental results on real data show that the proposed technique provides improved performance over naive methods.

• Smart Media Journal
• /
• v.13 no.4
• /
• pp.23-32
• /
• 2024

Large-scale fires and their consequential damages are becoming increasingly common, but confidence in fire detection systems is waning. Recently, widely-used chemical fire detectors frequently generate lots of false alarms, while video-based deep learning fire detection is hampered by its time-consuming and expensive nature. To tackle these issues, this study proposes a fire detection model utilizing an autoencoder approach. The objective is to minimize false alarms while achieving swift and precise fire detection. The proposed model, employing an autoencoder methodology, can exclusively learn from normal data without the need for fire-related data, thus enhancing its adaptability to diverse environments. By amalgamating data from five distinct sensors, it facilitates rapid and accurate fire detection. Through experiments with various hyperparameter combinations, the proposed model demonstrated that out of 14 scenarios, only one encountered false alarm issues. Experimental results underscore its potential to curtail fire-related losses and bolster the reliability of fire detection systems.

• Asian Journal of Innovation and Policy
• /
• v.13 no.1
• /
• pp.1-28
• /
• 2024

Enhancing the efficiency of research and development (R&D) is crucial for organizations to remain competitive and generate innovative solutions. Data Envelopment Analysis (DEA) has emerged as a powerful tool for evaluating R&D efficiency. However, traditional DEA models heavily rely on the selection of input and output variables, which can limit their effectiveness. To overcome this dependency and improve the robustness of DEA, this study proposes a novel methodology that integrates machine learning techniques with DEA for determining the most suitable input and output variables. The proposed approach is particularly relevant for specialized R&D fields, such as Radiation Emergency Medicine (REM). REM is a critical domain that deals with the medical and public health consequences of nuclear emergencies. The selection of REM as the focus of this study is motivated by several factors, including the unique challenges posed by the field, the potential for significant societal impact, and the need for efficient resource allocation in emergency situations. By leveraging machine learning algorithms, such as Support Vector Machines (SVM), the proposed methodology aims to identify the most relevant input and output variables for DEA in the context of REM. The integration of machine learning enables the DEA model to capture complex relationships and non-linearities in the data, leading to more accurate and reliable efficiency assessments. The effectiveness of the proposed methodology is demonstrated through a comprehensive evaluation using real-world REM data. The results highlight the superior performance of the machine learning-integrated DEA approach compared to traditional DEA models. This study contributes to the advancement of R&D efficiency assessment in specialized fields and provides valuable insights for decision-makers in REM and other critical domains.

• Journal of Korea Multimedia Society
• /
• v.17 no.9
• /
• pp.1115-1124
• /
• 2014

The objective of this study is to present a new design method at the idea brainstorming process, which requires a great deal of time and effort to seek their own ideas in existing design process and to develop a user support idea forms generation system called OCIGS (Openware Creative Idea Generation System) that makes it possible to create a variety and 3D forms as well as a quantitative evaluation data for the actual types that compose the aesthetic factors and designs of qualitative standards. In order to achieve these objectives : First, overcoming the limitation of 3D form generation ability by establishing visual application system, Second, accepting users' opinions by generating 3D images dynamically, analysing and giving information on the preferred ones on the web on real time, Third, identifying tendency of preference so that they can generate preferred colors and images in future by updating image combination and dropping low-preferred ones. This system would play a role as an idea or form generation application in the actual 3d form design development process.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2003.10a
• /
• pp.205-209
• /
• 2003

The urbanization of Korea has been rapidly progressed since 1960. Current available satellite images used in various fields are obtained after 1975. The CORONA Image data declassified in 1995, and are the only source of image which provide 1960's topographic information of the Korean Peninsula. In this sense CORONA imagery can be readily applicable for change detection in various fields such as urban, forest and environmental planning. To generate CORONA mosaic image of Korea we undertook comparative analysis of the various geocoding methods. We also applied the linear regression method to perform the radiometric balance between the strips.