• Journal of the Korea Computer Graphics Society
• /
• v.27 no.5
• /
• pp.63-71
• /
• 2021

Among drone autonomous flight technologies, obstacle avoidance is a very important technology that can prevent damage to drones or surrounding environments and prevent danger. Although the LiDAR sensor-based obstacle avoidance method shows relatively high accuracy and is widely used in recent studies, it has disadvantages of high unit price and limited processing capacity for visual information. Therefore, this paper proposes an obstacle avoidance algorithm for drones using camera-based PPO(Proximal Policy Optimization) reinforcement learning, which is relatively inexpensive and highly scalable using visual information. Drone, obstacles, target points, etc. are randomly located in a learning environment in the three-dimensional space, stereo images are obtained using a Unity camera, and then YOLov4Tiny object detection is performed. Next, the distance between the drone and the detected object is measured through triangulation of the stereo camera. Based on this distance, the presence or absence of obstacles is determined. Penalties are set if they are obstacles and rewards are given if they are target points. The experimennt of this method shows that a camera-based obstacle avoidance algorithm can be a sufficiently similar level of accuracy and average target point arrival time compared to a LiDAR-based obstacle avoidance algorithm, so it is highly likely to be used.

• Smart Media Journal
• /
• v.11 no.5
• /
• pp.9-16
• /
• 2022

Recently, agricultural sites are automating into digital agricultural smart farms by applying technologies such as big data and Internet of Things (IoT). These smart farms aim to increase production and improve crop quality by measuring the environment of crops, investigating and processing data. Production prediction is an important study in smart farm digital agriculture, which is a high-tech agriculture, and it is necessary to analyze environmental data using big data and further standardized research to manage the quality of growth information data. In this paper, environmental and production data collected from smart farm strawberry farms were analyzed and studied. Based on regression analysis, crop production prediction models were analyzed using Ridge Regression, LightGBM, and XGBoost. Among the three models, the optimal model was XGBoost, and R2 showed 82.5 percent explanatory power. As a result of the study, the correlation between the amount of positive fluid absorption and environmental data was confirmed, and significant results were obtained for the production prediction study. In the future, it is expected to contribute to the prevention of environmental pollution and reduction of sheep through the management of sheep by studying the amount of sheep absorption, such as information on the growing environment of crops and the ingredients of sheep.

• Journal of Korea Entertainment Industry Association
• /
• v.15 no.2
• /
• pp.123-135
• /
• 2021

The data processing of this study is focused on keywords related to 'Corona 19 and professional baseball' and 'Corona 19 and professional baseball no spectators', using text mining and social network analysis of textom program to identify problems and view quality. It was used to set the variable of For quantitative analysis, a questionnaire on viewing quality was constructed, and out of 270 survey respondents, 250 questionnaires were used for the final study. As a tool for securing the validity and reliability of the questionnaire, exploratory factor analysis and reliability analysis were conducted, and IPA analysis (importance-satisfaction) was conducted based on the questionnaire that secured validity and reliability, and the results and strategies were presented. As a result of IPA analysis, factors related to the image (image composition, image coloration, image clarity, image enlargement and composition, high-quality image) were found in the first quadrant, and the second quadrant was the game situation (support team game level, support player game level, star). Player discovery, competition with rival teams), game information (match schedule information, player information check, team performance and player performance, game information), interaction (consensus with the supporting team), and some factors appeared. The factors of commentator (baseball-related knowledge, communication ability, pronunciation and voice, use of standard language, introduction of game-related information) and interaction (real-time communication with the front desk, sympathy with viewers, information exchange such as chatting) appeared.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_1
• /
• pp.1015-1023
• /
• 2022

The Entry-Descent-Landing process of a lander involves many environmental and technical challenges. To solve these problems, recently, terrestrial relative navigation (TRN) technology has been essential for landers. TRN is a technology for estimating the position and attitude of a lander by comparing Inertial Measurement Unit (IMU) data and image data collected from a descending lander with pre-built reference data. In this paper, we present a method for generating descent dataset and extracting landmarks, which are key elements for developing TRN technologies to be used on Mars. The proposed method generates IMU data of a descending lander using a simulated Mars landing trajectory and generates descent images from high-resolution ortho-map and digital elevation map through a ray tracing technique. Landmark extraction is performed by an area-based extraction method due to the low-textured surfaces on Mars. In addition, search area reduction is carried out to improve matching accuracy and speed. The performance evaluation result for the descent dataset generation method showed that the proposed method can generate images that satisfy the imaging geometry. The performance evaluation result for the landmark extraction method showed that the proposed method ensures several meters of positioning accuracy while ensuring processing speed as fast as the feature-based methods.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.26 no.4
• /
• pp.57-64
• /
• 2022

Chloride, which is one of the main deterioration factors in reinforced concrete structures, can degrade the performance of the structure due to chloride-induced corrosion of steel. Chloride content at steel depth or the rate of chloride penetration is necessary to determine deterioration of reinforced concrete or to calculate initiation time of steel corrosion caused by chloride attack. Chlorides in concrete are generally identified with typical two methods including chloride profiling using potentiometric titration method and discoloration method using AgNO₃ solution. The former is advantageous to estimate chloride penetration rate (diffusion coefficient in general) with measured chloride contents directly, but it is laborious. In the case of latter, while the result is obtained easily with the range of discoloration, the error may occur depending on workmanship when the depth of chloride ingress is measured. This study shows that chloride penetrated depth is evaluated with the results obtained from discoloration method through image analysis, thereby the error is minimized by workmanship. In addition, the effect of micro-crack in concrete is studied on chloride penetration. In conclusion, the depth of chloride penetration was quantified with image analysis and as it was confirmed that chlorides can rapidly penetrate through micro-cracks, caution is especially required for cracks in concrete structure.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.10
• /
• pp.323-332
• /
• 2022

As various services using AI technology are being developed, much attention is being paid to AI service production. Recently, AI technology is acknowledged as one of ICT services, a lot of research is being conducted for general-purpose AI service production. In this paper, I describe the research results in terms of systems for AI service production, focusing on the distribution and production of machine learning models, which are the final steps of general machine learning development procedures. Three different Ubuntu systems were built, and experiments were conducted on the system, using data from 2017 validation COCO dataset in combination of different AI models (RFCN, SSD-Mobilenet) and different communication methods (gRPC, REST) to request and perform AI services through Tensorflow serving. Through various experiments, it was found that the type of AI model has a greater influence on AI service inference time than AI machine communication method, and in the case of object detection AI service, the number and complexity of objects in the image are more affected than the file size of the image to be detected. In addition, it was confirmed that if the AI service is performed remotely rather than locally, even if it is a machine with good performance, it takes more time to infer the AI service than if it is performed locally. Through the results of this study, it is expected that system design suitable for service goals, AI model development, and efficient AI service production will be possible.

• Analytical Science and Technology
• /
• v.35 no.5
• /
• pp.205-211
• /
• 2022

Deuterium (D) is an isotope with one more neutron number than hydrogen (H). Heavy elements rarely change their chemical properties with little effect even if the number of neutrons increases, but low-mass elements change their vibration energy, diffusion rate, and reaction rate because the effect cannot be ignored, which is called an isotope effect. Recently, in the semiconductor and display industries, there is a trend to replace hydrogen gas (H₂) with deuterium gas (D₂) in order to improve process stability and product quality by using the isotope effect. In addition, as the demand for D₂ in industries increases, domestic gas producers are making efforts to produce and supply D₂ on their own. In the case of high purity D₂, most of them are produced by electrolysis of heavy water (D₂O), and among D₂, hydrogen deuteride (HD) molecules are present as isotope impurities. Therefore, in order to maximize the isotope effect of hydrogen in the electronic industry, HD, which is an isotope impurity of D₂ used in the process, should be small amount. To this end, purity analysis of D₂ for industrial processing is essential. In this study, HD quantitative analysis of D₂ for high purity D₂ purity analysis was established and hydrogen isotope RM (Reference material) was developed. Since hydrogen isotopes are difficult to analyze with general gas analysis instrument, they were analyzed using a high-precision mass spectrometer (Gas/MS, Finnigan MAT271). High purity HD gas was injected into Gas/MS, sensitivity was determined by a signal according to pressure, and HD concentrations in two bottles of D₂ were quantified using the corresponding sensitivity. The amount fraction of HD in each D₂ was (4518 ± 275) μmol/mol, (2282 ± 144) μmol/mol. D₂, which quantifies HD amount using the developed quantitative analysis method, will be manufactured with hydrogen isotope RM and distributed for quality management and maintenance of electronic industries and gas producers in the future.

• Journal of Korea Entertainment Industry Association
• /
• v.14 no.7
• /
• pp.79-94
• /
• 2020

The purpose of this study is to analyze the difference in the effect of presence for the activation of virtual reality sports on participation satisfaction, to suggest continuous screen golf exercise participation, and to provide empirical and academic data for the development of the entire virtual reality sports market. To achieve this purpose, the survey period was from March 13 to May 13, 2020, with five researchers and assistants. The purpose of this study and the questionnaire were fully explained to consumers who experienced screen golf directly, and 247 questionnaires were used as the final valid sample by making a questionnaire with self-administration method. The data processing method was the statistical program Windows SPSS 18.0. First, factor analysis and reliability analysis, second, frequency analysis mean(M) and standard deviation(SD), third, Scheffe analysis among t-test and One-way ANOVA analysis, fourth, correlation analysis between variables and multiple regression analysis were conducted. The results of this study through these methods and procedures are as follows. First, there was a significant difference in participation satisfaction of presence in gender, and participation period of general characteristics. Second, there was a high difference in social presence, social self-reliance, and Ego, which are sub-factors of Presence, in social satisfaction, psychological satisfaction, and physical satisfaction. Third, the sub-factors of Presence, Social Presence, Social Self-Reliance, and Ego, were found to have a high effect on the sub-factors of Participation Satisfaction, Social Satisfaction, and Psychological Satisfaction.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.3
• /
• pp.25-33
• /
• 2023

Noise generated during the acquisition and transmission of CT images acts as a factor that degrades image quality. Therefore, noise removal to solve this problem is an important preprocessing process in image processing. In this paper, we remove noise by using a deformable convolutional autoencoder (DeCAE) model in which deformable convolution operation is applied instead of the existing convolution operation in the convolutional autoencoder (CAE) model of deep learning. Here, the deformable convolution operation can extract features of an image in a more flexible area than the conventional convolution operation. The proposed DeCAE model has the same encoder-decoder structure as the existing CAE model, but the encoder is composed of deformable convolutional layers and the decoder is composed of conventional convolutional layers for efficient noise removal. To evaluate the performance of the DeCAE model proposed in this paper, experiments were conducted on CT images corrupted by various noises, that is, Gaussian noise, impulse noise, and Poisson noise. As a result of the performance experiment, the DeCAE model has more qualitative and quantitative measures than the traditional filters, that is, the Mean filter, Median filter, Bilateral filter and NL-means method, as well as the existing CAE models, that is, MAE (Mean Absolute Error), PSNR (Peak Signal-to-Noise Ratio) and SSIM. (Structural Similarity Index Measure) showed excellent results.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.3
• /
• pp.91-100
• /
• 2023

This study is a case study comparing and examining the effects of non-face-to-face(NFTF) classes in the 2021-2 semester and face-to-face(FTF) classes in the 2022-2 semester on learning immersion, learning stress, and learning satisfaction. The learning immersion and learning satisfaction of 240 students were analyzed in NFTF and FTF classes of department S of C junior college where the same textbook, same subject, and same professor were taught. For data processing, SPSS Ver. 23.0 was used. The data is used to measure reliability by Cronbach's α, t-test, Pearson's correlation coefficient, and multiple regression analysis. The results of this study are as follows. First, learners' learning immersion was higher in FTF than NFTF classes among engineering major subjects. Second, it was found that there was a difference in learning stress according to the types of FTF and NFTF classes in engineering major subjects. Third, it was found that there were differences in practice content, communication, and task performance of sub-factors of learning satisfaction according to FTF and NFTF class types in engineering major subjects. In conclusion, it was found that FTF classes had a more positive effect on learning immersion and satisfaction, and NFTF classes had a more negative effect on learning stress.