• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.153-153
• /
• 2022

Large-scale and accurate observations at fine spatial resolution through a means of remote sensing offer an effective tool for capturing rainfall variability over the traditional rain gauges and weather radars. Although satellite rainfall products (SRPs) derived using two major estimation approaches were evaluated worldwide, their practical applications suffered from limitations. In particular, the traditional top-down SRPs (e.g., IMERG), which are based on direct estimation of rain rate from microwave satellite observations, are mainly restricted with their coarse spatial resolution, while applications of the bottom-up approach, which allows backward estimation of rainfall from soil moisture signals, to novel high spatial resolution soil moisture satellite sensors over South Korea are not introduced. Thus, this study aims to evaluate the performances of a state-of-the-art bottom-up SRP (the self-calibrated SM2RAIN model) applied to the C-band SAR Sentinel-1, a statistically downscaled version of the conventional top-down IMERG SRP, and their integration for a targeted high spatial resolution of 0.01° (~ 1-km) over central South Korea, where the differences in climate zones (coastal region vs. mainland region) and vegetation covers (croplands vs. mixed forests) are highlighted. The results indicated that each single SRP can provide plus points in distinct climatic and vegetated conditions, while their drawbacks have existed. Superior performance was obtained by merging these individual SRPs, providing preliminary results on a complementary high spatial resolution SRP over central South Korea. This study results shed light on the further development of integration framework and a complementary high spatial resolution rainfall product from multi-satellite sensors as well as multi-observing systems (integrated gauge-radar-satellite) extending for entire South Korea, toward the demands for urban hydrology and microscale agriculture.

• Asia Marketing Journal
• /
• v.14 no.3
• /
• pp.7-26
• /
• 2012

The success story of E-mart fascinated many academics and practitioners alike. Though E-mart began as a nameless discount store in Chang-dong, Seoul in 1993, it has transformed itself into a leading distribution company and one of the most powerful brands in Korea. Surprisingly, it achieved the great success against the two crises it met: the national economic crisis and the invasion of the global giant Walmart. The main objective of this case study is to formally examine how E-mart overcame the two crises. More specifically, this case study highlights the ways with which E-mart turned those difficulties into opportunities for growth. In our examination of the E-mart case, we could clearly see E-mart's competence and spirit that allowed it to turn crises into advantageous opportunities. E-mart attracted the customers who wanted value-oriented consumption by its positioning as the "Lowest price discount store", when consumer sentiment was frozen under the economic crisis. Furthermore, when a large-scale foreign discount store like Walmart entered the Korea market, E-mart built its core competencies as the 'Korean style discount store'. These ingenious positioning and efforts resulted in E-mart taking over their archrival, Walmart, and forced the global Goliath to exit the Korean market. The case of E-mart's effective crisis management teaches many important lessons and a few core lessons that apply to many companies. One such lesson is the importance of positioning which enabled E-mart to turn crises into opportunities. Granted, the strategy of positioning as the 'Korean style discount store', or 'Lowest price discount store' was possible due to overall support with cost reduction, development and management of their own system, an apprentice educate system, etc. based on an excellent selection of location of the store and efficient distribution systems. Still, the positioning strategy of E-mart was truly ground breaking in distancing itself from its competitors. The lessons from E-mart will help those companies currently in a stagnant situation or a crisis to turn their obstacles into great success.

• Journal of Powder Materials
• /
• v.30 no.6
• /
• pp.516-520
• /
• 2023

Highly safe lithium-ion batteries (LIBs) are required for large-scale applications such as electrical vehicles and energy storage systems. A highly stable cathode is essential for the development of safe LIBs. LiFePO₄ is one of the most stable cathodes because of its stable structure and strong bonding between P and O. However, it has a lower energy density than lithium transition metal oxides. To investigate the high energy density of phosphate materials, vanadium phosphates were investigated. Vanadium enables multiple redox reactions as well as high redox potentials. LiVPO₄O has two redox reactions (V⁵⁺/V⁴⁺/V³⁺) but low electrochemical activity. In this study, LiVPO₄O is doped with fluorine to improve its electrochemical activity and increase its operational redox potential. With increasing fluorine content in LiVPO₄O_1-xF_x, the local vanadium structure changed as the vanadium oxidation state changed. In addition, the operating potential increased with increasing fluorine content. Thus, it was confirmed that fluorine doping leads to a strong inductive effect and high operating voltage, which helps improve the energy density of the cathode materials.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.46 no.4
• /
• pp.15-31
• /
• 2023

In this study, we propose a novel approach to analyze big data related to patents in the field of smart factories, utilizing the Latent Dirichlet Allocation (LDA) topic modeling method and the generative artificial intelligence technology, ChatGPT. Our method includes extracting valuable insights from a large data-set of associated patents using LDA to identify latent topics and their corresponding patent documents. Additionally, we validate the suitability of the topics generated using generative AI technology and review the results with domain experts. We also employ the powerful big data analysis tool, KNIME, to preprocess and visualize the patent data, facilitating a better understanding of the global patent landscape and enabling a comparative analysis with the domestic patent environment. In order to explore quantitative and qualitative comparative advantages at this juncture, we have selected six indicators for conducting a quantitative analysis. Consequently, our approach allows us to explore the distinctive characteristics and investment directions of individual countries in the context of research and development and commercialization, based on a global-scale patent analysis in the field of smart factories. We anticipate that our findings, based on the analysis of global patent data in the field of smart factories, will serve as vital guidance for determining individual countries' directions in research and development investment. Furthermore, we propose a novel utilization of GhatGPT as a tool for validating the suitability of selected topics for policy makers who must choose topics across various scientific and technological domains.

• Journal of Astronomy and Space Sciences
• /
• v.40 no.4
• /
• pp.173-197
• /
• 2023

ShadowCam is a high-sensitivity, high-resolution imager provided by NASA for the Danuri (KPLO) lunar mission. ShadowCam calibration shows that it is well suited for its purpose, to image permanently shadowed regions (PSRs) that occur near the lunar poles. It is 205 times as sensitive as the Lunar Reconnaissance Orbiter Camera (LROC) Narrow Angle Camera (NAC). The signal to noise ratio (SNR) is greater than 100 over a large part of the dynamic range, and the top of the dynamic range is high enough to accommodate most brighter PSR pixels. The optical performance is good enough to take full advantage of the 1.7 meter/pixel image scale, and calibrated images have uniform response. We describe some instrument artifacts that are amenable to future corrections, making it possible to improve performance further. Stray light control is very challenging for this mission. In many cases, ShadowCam can image shadowed areas with directly illuminated terrain in or near the field of view (FOV). We include thorough qualitative descriptions of circumstances under which lunar brightness levels far higher than the top of the dynamic range cause detector or stray light artifacts and the size and extent of the artifact signal under those circumstances.

• Journal of the Korean Institute of Gas
• /
• v.28 no.1
• /
• pp.53-58
• /
• 2024

Due to the development of industry and improvement of living standards, the amount of natural gas used in the world is constantly increasing, and related industrial facilities such as power plants, storage facilities, and supply pipelines are constantly increasing. Natural gas is a convenient and clean fuel that does not pollute the environment, but in the event of an accident due to leakage, it can cause human casualties, large-scale property damage, and negative effects on the global warming effect. In addition to the severe penalties under the Severe Disaster Punishment Act, it is necessary to ensure safety. Therefore, by applying the principle of laser-based absorption spectroscopy, we developed a long-range portable methane leakage gas detection system that can detect the concentration of methane leaking from a distance of up to 30 meters and verified its effectiveness.

• Journal of Information Processing Systems
• /
• v.20 no.3
• /
• pp.375-390
• /
• 2024

Edge computing architecture has effectively alleviated the computing pressure on cloud platforms, reduced network bandwidth consumption, and improved the quality of service for user experience; however, it has also introduced new security issues. Existing anomaly detection methods in big data scenarios with cloud-edge computing collaboration face several challenges, such as sample imbalance, difficulty in dealing with complex network traffic attacks, and difficulty in effectively training large-scale data or overly complex deep-learning network models. A lightweight deep-learning model was proposed to address these challenges. First, normalization on the user side was used to preprocess the traffic data. On the edge side, a trained Wasserstein generative adversarial network (WGAN) was used to supplement the data samples, which effectively alleviates the imbalance issue of a few types of samples while occupying a small amount of edge-computing resources. Finally, a trained lightweight deep learning network model is deployed on the edge side, and the preprocessed and expanded local data are used to fine-tune the trained model. This ensures that the data of each edge node are more consistent with the local characteristics, effectively improving the system's detection ability. In the designed lightweight deep learning network model, two sets of convolutional pooling layers of convolutional neural networks (CNN) were used to extract spatial features. The bidirectional long short-term memory network (BiLSTM) was used to collect time sequence features, and the weight of traffic features was adjusted through the attention mechanism, improving the model's ability to identify abnormal traffic features. The proposed model was experimentally demonstrated using the NSL-KDD, UNSW-NB15, and CIC-ISD2018 datasets. The accuracies of the proposed model on the three datasets were as high as 0.974, 0.925, and 0.953, respectively, showing superior accuracy to other comparative models. The proposed lightweight deep learning network model has good application prospects for anomaly traffic detection in cloud-edge collaborative computing architectures.

• Fisheries and Aquatic Sciences
• /
• v.27 no.5
• /
• pp.265-275
• /
• 2024

The success of establishing the Indonesian growing fast hybrid carp, namely "Permata", on a controlled environmental test must be followed up with a large-scale test. This study aims to evaluate the phenotypic performance of the Permata hybrid carp in multi-locations with different cultivation systems. The test sites consisted of floating net cages, running-water ponds, semi-concrete ponds, earthen ponds, fully concrete ponds, and static net cages. For 90 days, fish were fed commercial pellets with a 28%-30% protein content. At the end of the test, all fish were harvested and counted. Data on length, weight, survival rate, and harvested biomass were used to analyze the effect of genotype, environment, and their interaction on the phenotypic performance. The growth based on final weight is used to analyze the stability performance in each test location. The results showed that the length and weight of common carp were significantly affected by genotype and the environment, but not by the interaction of both. The genotype, environment, and the interaction of both factors affected common carp's survival and harvested biomass. Common carp reared in floating net cages generally had the best performance, while carp reared in fully concrete tanks and static net cages had the lowest. The growth stability analysis showed that the common carp in this study were unstable genotypes but have a broad adaptability in term of different environments.

• Journal of Practical Agriculture & Fisheries Research
• /
• v.26 no.3
• /
• pp.45-52
• /
• 2024

In this study, we analyzed the working environment of wholesale market distribution workers and suggested policy directions for the maintenance, operation, and development of sustainable wholesale market functions in response to changes in agricultural product distribution environment and working environment. The results of the analysis showed that there is a large gap in the working environment between wholesale corporation workers and middle wholesalers, and overall, the level of satisfaction with the working environment was low. In order to maintain sustainable wholesale market functions in the future, various policies and support should be established to improve the working environment of wholesale market distribution workers. The directions are as follows. First, wholesale market corporations should expand support for improving working environments by introducing systems such as rotational work by investing a portion of their profits in expanding manpower in response to changes in agricultural product distribution environment and working environment. Some Corporation of Garak-dong Wholesale Market is making great efforts to improve the working environment by introducing a rotational work system by expanding the workforce of auctioneers and providing a practical 5-day work week. In addition, in the case of wholesalers, it is expected that most of them will have difficulty in increasing their workforce as they are small businesses. However, it is judged that consolidation of businesses among wholesalers can be an alternative to increasing their workforce through expanding their management scale.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.