• Korean Journal of Agricultural and Forest Meteorology
• /
• v.21 no.2
• /
• pp.97-109
• /
• 2019

Degree days have been determined using temperature data measured at nearby weather stations to a site of interest to produce information for supporting decision-making on agricultural production. Alternatively, the data products of Moderate Resolution Imaging Spectroradiometer (MODIS) can be used for estimation of degree days in a given region, e.g., Korean Peninsula. The objective of this study was to develop a simple tool for processing the MODIS product for estimating cooling degree days (CDD), which would help assessment of heat stress conditions for a crop as well as energy requirement for greenhouses. A set of scripts written in R was implemented to obtain temperature profile data for the region of interest. These scripts had functionalities for processing spatial data, which include reprojection, mosaicking, and cropping. A module to extract air temperature at the surface pressure level was also developed using R extension packages such as rgdal and RcppArmadillo. Random forest (RF) models, which estimate mean temperature and CDD with a different set of MODIS data, were trained at 34 sites in South Korea during 2009 - 2018. Then, the values of CDD were calculated over Korean peninsula during the same period using those RF models. It was found that the CDD estimates using the MODIS data explained >74% of the variation in the CDD measurements at the weather stations in North Korea as well as South Korea. These results indicate that temperature data derived from the MODIS atmospheric products would be useful for reliable estimation of CDD. Our results also suggest that the MODIS data can be used for preparation of weather input data for other temperature-based agro-ecological models such as growing degree days or chill units.

• The Journal of Bigdata
• /
• v.3 no.2
• /
• pp.59-70
• /
• 2018

In order to provide stable district heat supplying service to the certain limited residential area, it is the most important to forecast the short-term future demand more accurately and produce and supply heat in efficient way. However, it is very difficult to develop a universal heat demand forecasting model that can be applied to general situations because the factors affecting the heat consumption are very diverse and the consumption patterns are changed according to individual consumers and regional characteristics. In particular, considering all of the various variables that can affect heat demand does not help improve performance in terms of accuracy and versatility. Therefore, this study aims to develop a demand forecasting model using deep learning based on only limited information that can be acquired in real time. A demand forecasting model was developed by learning the artificial neural network of the Tensorflow using past data consisting only of the outdoor temperature of the area and date as input variables. The performance of the proposed model was evaluated by comparing the accuracy of demand predicted with the previous regression model. The proposed heat demand forecasting model in this research showed that it is possible to enhance the accuracy using only limited variables which can be secured in real time. For the demand forecasting in a certain region, the proposed model can be customized by adding some features which can reflect the regional characteristics.

• Journal of Korean Society on Water Environment
• /
• v.35 no.3
• /
• pp.224-233
• /
• 2019

Nitrogen components in liquid manure can reduce safety and quality of environment harmfully. To minimize the environmental risks of manure, understanding fate of manure in environment is necessary. This study aimed at investigating applicability of a simplified Level III fugacity model for simulating $NH_3-N$ component to analyze environmental fate and transport of $NH_3-N$ in liquid manure and to provide basis for improving management of N in the liquid manure system and for minimizing the environmental impacts of N. The model simulation conducted for four environmental compartments (air, water, soil, and rice plants) during rice-cropping to trace $NH_3-N$ component and provided applicability of the Level III fugacity model in studying the environmental fate of $NH_3-N$ in manure. Most of $NH_3-N$ was found in water body and in rice plants depending upon the physicochemical properties and proper removal processes. For more precise model results, the model is needed to modify with the detailed removal processes in each compartment and to collect proper and accurate information for input parameters. Further study should be about simulations of various N-typed fertilizers to compare with the liquid manure based on a modified and relatively simplified Level III fugacity model.

• Journal of the Korea Computer Graphics Society
• /
• v.27 no.3
• /
• pp.65-74
• /
• 2021

3D reconstruction means that reconstructing the 3D shape of the object in an image and a video. We proposed a progressive occupancy network architecture that can recover not only the overall shape of the object but also the local details. Unlike the original occupancy network, which uses a feature vector embedding information of the whole image, we extract and utilize the different levels of image features depending on the receptive field size. We also propose a novel network architecture that applies the image features sequentially to the decoder blocks in the decoder and improves the quality of the reconstructed 3D shape progressively. In addition, we design a novel decoder block structure that combines the different levels of image features properly and uses them for updating the input point feature. We trained our progressive occupancy network with ShapeNet. We compare its representation power with two prior methods, including prior occupancy network(ONet) and the recent work(DISN) that used different levels of image features like ours. From the perspective of evaluation metrics, our network shows better performance than ONet for all the metrics, and it achieved a little better or a compatible score with DISN. For visualization results, we found that our method successfully reconstructs the local details that ONet misses. Also, compare with DISN that fails to reconstruct the thin parts or occluded parts of the object, our progressive occupancy network successfully catches the parts. These results validate the usefulness of the proposed network architecture.

• Journal of the Korean Society of Radiology
• /
• v.14 no.7
• /
• pp.991-1001
• /
• 2020

Supervised deep learning technologies for improving the image quality of computed tomography (CT) need a lot of training data. When input images have different characteristics with training images, the technologies cause structural distortion in output images. In this study, an imaging model based on the deep reinforcement learning (DRL) was developed for overcoming the drawbacks of the supervised deep learning technologies and reducing noise in CT images. The DRL model was consisted of shared, value and policy networks, and the networks included convolutional layers, rectified linear unit (ReLU), dilation factors and gate rotation unit (GRU) in order to extract noise features from CT images and improve the performance of the DRL model. Also, the quality of the CT images obtained by using the DRL model was compared to that obtained by using the supervised deep learning model. The results showed that the image accuracy for the DRL model was higher than that for the supervised deep learning model, and the image noise for the DRL model was smaller than that for the supervised deep learning model. Also, the DRL model reduced the noise of the CT images, which had different characteristics with training images. Therefore, the DRL model is able to reduce image noise as well as maintain the structural information of CT images.

• Journal of the Korea Convergence Society
• /
• v.12 no.2
• /
• pp.45-50
• /
• 2021

In this paper, we propose the development of a smart space security device that can be opened and closed remotely using IoT. Existing space security devices can control opening and closing by breaking hardware or only using button devices or replicated keys. The recent COVID-19 crisis has created several applications for non-contact devices. In this study, we propose the development of a small space security device that has the function of unlocking through an app without touching the device. By transferring the control authority to a smartphone, device that cannot be opened or closed by only operating hardware at the user's option. It is convenient and hygienic because it can be opened and closed using an app without touching the locking device. Multiple security is possible because security can be released using an app after user authentication by fingerprint recognition and pattern input on a smartphone. If the user wishes, after using the app security, the security is released by directly touching a button installed in the safe or space or opening it with a key. In addition, by adding an inactive function to the app, it is designed so that the door of the safe cannot be opened when the key is lost or the small safe is lost. This study is expected to be able to effectively expand the security system by applying variously to objects that require security.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.6
• /
• pp.116-126
• /
• 2021

Ground Component Command (GCC) has been developing operational planning and execution systems to implement "Decisive Integrated Operations", which is the concept of ground operations execution, and achieved remarkable results. In particular, "Simultaneous Offense-Defense Integrated Operations" is developed mainly to neutralize enemies in deep areas and develop favorable conditions for the allies early by simultaneously attacking and defending from the beginning of the war. On the other hand, it is limited to providing scientific and reasonable support for the commander's decision-making process because analyzing the effects of the deep operation with existing M&S systems is impossible. This study developed a model for analyzing the effects of deep operations that can be used in the KJCCS. Previous research was conducted on the effects of surveillance, physical strike, and non-physical strike, which are components of deep operations to find the characteristics and limitations and suggest a research direction. A methodology for analyzing the effects of deep operations reflecting the interactions of components using data was then developed by the GCC, and input data for each field was calculated through combat experiments and a literature review. Finally, the Deep operations Effect CAlculating Model(DECAM) was developed and distributed to the GCC and Corps battle staff during the ROK-US Combined Exercise. Through this study, the effectiveness of the methodology and the developed model were confirmed and contribute to the development of the GCC and Corps' abilities to perform deep operations.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.40 no.6
• /
• pp.613-622
• /
• 2020

The South Korean government plans for a 37 % reduction in CO2 emissions against business as usual by 2030. Subsequently, the Ministry of Land, Infrastructure and Transport declared a 26.9 % reduction target in greenhouse gas emissions from buildings by 2020 and established the Green Standard for Energy and Environmental Design (G-SEED) to help improve the environmental performance of buildings. Construction companies often work with consulting firms to prepare for G-SEED certification. In the process, owing to inefficient data sharing and work connections, it is difficult to achieve economic efficiency and obtain certification. The objective of this study was to develop an economic model to assist contractors in achieving the required G-SEED scores for materials and resources. To do this, we automated the process for material comparison and selection on the basis of an analysis of actual consulting data, and developed a model that selects material alternatives that can meet the required scores at a minimum cost. Information on materials is input by applying a genetic algorithm to the optimization of alternatives. When the model was applied to actual data, the construction cost could be lowered by 79.3 % compared with existing methods. The economical material selection model is expected to not only reduce construction costs for owners desiring G-SEED certification but also shorten the project design time.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.3
• /
• pp.141-148
• /
• 2022

Recently, reinforcement learning combined with deep neural network technology has achieved remarkable success in various fields such as board games such as Go and chess, computer games such as Atari and StartCraft, and robot object manipulation tasks. However, such deep reinforcement learning describes states, actions, and policies in vector representation. Therefore, the existing deep reinforcement learning has some limitations in generality and interpretability of the learned policy, and it is difficult to effectively incorporate domain knowledge into policy learning. On the other hand, dNL-RRL, a new relational reinforcement learning framework proposed to solve these problems, uses a kind of vector representation for sensor input data and lower-level motion control as in the existing deep reinforcement learning. However, for states, actions, and learned policies, It uses a relational representation with logic predicates and rules. In this paper, we present dNL-RRL-based policy learning for transportation mobile robots in a manufacturing environment. In particular, this study proposes a effective method to utilize the prior domain knowledge of human experts to improve the efficiency of relational reinforcement learning. Through various experiments, we demonstrate the performance improvement of the relational reinforcement learning by using domain knowledge as proposed in this paper.

• Knowledge Management Research
• /
• v.22 no.4
• /
• pp.27-40
• /
• 2021

Click-Through Rate (CTR) prediction is a key function that determines the ranking of candidate items in the recommendation system and recommends high-ranking items to reduce customer information overload and achieve profit maximization through sales promotion. The fields of natural language processing and image classification are achieving remarkable growth through the use of deep neural networks. Recently, a transformer model based on an attention mechanism, differentiated from the mainstream models in the fields of natural language processing and image classification, has been proposed to achieve state-of-the-art in this field. In this study, we present a method for improving the performance of a transformer model for CTR prediction. In order to analyze the effect of discrete and categorical CTR data characteristics different from natural language and image data on performance, experiments on embedding regularization and transformer normalization are performed. According to the experimental results, it was confirmed that the prediction performance of the transformer was significantly improved when the L2 generalization was applied in the embedding process for CTR data input processing and when batch normalization was applied instead of layer normalization, which is the default regularization method, to the transformer model.