• Korean Journal of Remote Sensing
• /
• v.37 no.2
• /
• pp.275-290
• /
• 2021

Sulfur dioxide (SO2) in the atmosphere is mainly generated from anthropogenic emission sources. It forms ultra-fine particulate matter through chemical reaction and has harmful effect on both the environment and human health. In particular, ground-level SO2 concentrations are closely related to human activities. Satellite observations such as TROPOMI (TROPOspheric Monitoring Instrument)-derived column density data can provide spatially continuous monitoring of ground-level SO2 concentrations. This study aims to propose a 2-step residual corrected model to estimate ground-level SO2 concentrations through the synergistic use of satellite data and numerical model output. Random forest machine learning was adopted in the 2-step residual corrected model. The proposed model was evaluated through three cross-validations (i.e., random, spatial and temporal). The results showed that the model produced slopes of 1.14-1.25, R values of 0.55-0.65, and relative root-mean-square-error of 58-63%, which were improved by 10% for slopes and 3% for R and rRMSE when compared to the model without residual correction. The model performance by country was slightly reduced in Japan, often resulting in overestimation, where the sample size was small, and the concentration level was relatively low. The spatial and temporal distributions of SO2 produced by the model agreed with those of the in-situ measurements, especially over Yangtze River Delta in China and Seoul Metropolitan Area in South Korea, which are highly dependent on the characteristics of anthropogenic emission sources. The model proposed in this study can be used for long-term monitoring of ground-level SO2 concentrations on both the spatial and temporal domains.

• Journal of Korea Water Resources Association
• /
• v.54 no.7
• /
• pp.535-544
• /
• 2021

Agricultural water occupies 48% of water demand, and management of agricultural reservoirs is essential for water resources management within agricultural basins. For more efficient use of agricultural water, monitoring the distribution of water resources in agricultural reservoirs and agricultural basins is required. Therefore, in this study, three threshold determination methods (i.e., fixed threshold, Otsu threshold, Kittler-Illingworth (KI) threshold) were compared to detect terrestrial water bodies using Sentinel-1 images for 3 years from 2018 to 2020. The purpose of this study was to evaluate methods for determining threshold values to more accurately estimate the reservoir area. In addition, by analyzing the relationship between the water surface and water storage at the Edong, Gosam, and Giheung reservoirs, water storage based on the SAR image was estimated and validated with observations. The thresholding method for detecting a waterbody was found to be the most accurate in the case of the KI threshold, and the water storage estimated by the KI threshold indicated a very high agreement (r = 0.9235, KGE' = 0.8691). Although the seasonal error characteristics were not observed, the problem of underestimation at high water levels may occur; the relationship between the water surface and the water storage could change rapidly. Therefore, it is necessary to understand the relationship between the water surface area and water storage through ground observation data for a more accurate estimation of water storage. If the use of SAR data through water resources satellites becomes possible in the future, based on the results of this study, it is judged that it will be beneficial for monitoring water storage and managing drought.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.22 no.3
• /
• pp.21-36
• /
• 2019

Dead fuel moisture content is a key variable in fire danger rating as it affects fire ignition and behavior. This study evaluates simple regression models estimating the moisture content of standardized 10-h fuel stick (10-h FMC) at three sites with different characteristics(urban and outside/inside the forest). Equilibrium moisture content (EMC) was used as an independent variable, and in-situ measured 10-h FMC was used as a dependent variable and validation data. 10-h FMC spatial distribution maps were created for dates with the most frequent fire occurrence during 2013-2018. Also, 10-h FMC values of the dates were analyzed to investigate under which 10-h FMC condition forest fire is likely to occur. As the results, fitted equations could explain considerable part of the variance in 10-h FMC (62~78%). Compared to the validation data, the models performed well with R2 ranged from 0.53 to 0.68, root mean squared error (RMSE) ranged from 2.52% to 3.43%, and bias ranged from -0.41% to 1.10%. When the 10-h FMC model fitted for one site was applied to the other sites, $R^2$ was maintained as the same while RMSE and bias increased up to 5.13% and 3.68%, respectively. The major deficiency of the 10-h FMC model was that it poorly caught the difference in the drying process after rainfall between 10-h FMC and EMC. From the analysis of 10-h FMC during the dates fire occurred, more than 70% of the fires occurred under a 10-h FMC condition of less than 10.5%. Overall, the present study suggested a simple model estimating 10-h FMC with acceptable performance. Applying the 10-h FMC model to the automatic mountain weather observation system was successfully tested to produce a national-scale 10-h FMC spatial distribution map. This data will be fundamental information for forest fire research, and will support the policy maker.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_3
• /
• pp.1827-1836
• /
• 2022

Disasters that occur unexpectedly are difficult to predict. In addition, the scale and damage are increasing compared to the past. Sometimes one disaster can develop into another disaster. Among the four stages of disaster management, search and rescue are carried out in the response stage when an emergency occurs. Therefore, personnel such as firefighters who are put into the scene are put in at a lot of risk. In this respect, in the initial response process at the disaster site, robots are a technology with high potential to reduce damage to human life and property. In addition, Light Detection And Ranging (LiDAR) can acquire a relatively wide range of 3D information using a laser. Due to its high accuracy and precision, it is a very useful sensor when considering the characteristics of a disaster site. Therefore, in this study, development and experiments were conducted so that the robot could perform real-time monitoring at the disaster site. Multi-sensor module was developed by combining LiDAR, Inertial Measurement Unit (IMU) sensor, and computing board. Then, this module was mounted on the robot, and a customized Simultaneous Localization and Mapping (SLAM) algorithm was developed. A method for stably mounting a multi-sensor module to a robot to maintain optimal accuracy at disaster sites was studied. And to check the performance of the module, SLAM was tested inside the disaster building, and various SLAM algorithms and distance comparisons were performed. As a result, PackSLAM developed in this study showed lower error compared to other algorithms, showing the possibility of application in disaster sites. In the future, in order to further enhance usability at disaster sites, various experiments will be conducted by establishing a rough terrain environment with many obstacles.

• Korean Journal of Remote Sensing
• /
• v.38 no.5_2
• /
• pp.747-763
• /
• 2022

Ocean salinity affects ocean circulation on a global scale and low salinity water around coastal areas often has an impact on aquaculture and fisheries. Microwave satellite sensors (e.g., Soil Moisture Active Passive [SMAP]) have provided sea surface salinity (SSS) based on the dielectric characteristics of water associated with SSS and sea surface temperature (SST). In this study, a Light Gradient Boosting Machine (LGBM)-based model for generating high resolution SSS from Geostationary Ocean Color Imager (GOCI) data was proposed, having machine learning-based improved SMAP SSS by Jang et al. (2022) as reference data (SMAP SSS (Jang)). Three schemes with different input variables were tested, and scheme 3 with all variables including Multi-scale Ultra-high Resolution SST yielded the best performance (coefficient of determination = 0.60, root mean square error = 0.91 psu). The proposed LGBM-based GOCI SSS had a similar spatiotemporal pattern with SMAP SSS (Jang), with much higher spatial resolution even in coastal areas, where SMAP SSS (Jang) was not available. In addition, when tested for the great flood occurred in Southern China in August 2020, GOCI SSS well simulated the spatial and temporal change of Changjiang Diluted Water. This research provided a potential that optical satellite data can be used to generate high resolution SSS associated with the improved microwave-based SSS especially in coastal areas.

• Journal of Korean Society of Forest Science
• /
• v.111 no.4
• /
• pp.622-629
• /
• 2022

The study aim was to derive a stem taper equation for Phyllostachys pubescens, a type of bamboo in South Korea, and to develop a stem volume table. To derive the stem taper equation, three stem taper models (Max & Burkhart, Kozak, and Lee) were used. Since bamboo stalks are hollow because of its woody characteristics, the outer and inner diameters of the tree were calculated, and connecting them enabled estimating the tree curves. The results of the three equations for estimating the outer and inner diameters led to selection of the Kozak model for determining the optimal stem taper because it had the highest fitness index and lowest error and bias. We used the Kozak model to estimate the diameter of Phyllostachys pubescens by stem height, which proved optimal, and drew the stem curve. After checking the residual degree in the stem taper equation, all residuals were distributed around "0", which proved the suitability of the equation. To calculate the stem volume of Phyllostachys pubescens, a rotating cube was created by rotating the stem curve with the outer diameter at 360°, and the volume was calculated by applying Smalian's method. The volume of Phyllostachys pubescens was calculated by deducting the inner diameter calculated volume from the outer diameter calculated volume. The volume of Phyllostachys pubescens was only 20~30% of the volume of Larix kaempferi, which is a general species. However, considering the current trees/ha of Phyllostachys pubescens and the amount of bamboo shoots generated every year, the individual tree volume was predicted to be small, but the volume/ha was not very different or perhaps more. The significance of this study is the stem taper equation and stem volume table for Phyllostachys pubescens developed for the first time in South Korea. The results are expected to be used as basic data for bamboo trading that is in increasing public and industrial demand and carbon absorption estimation.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.249-263
• /
• 2023

Collaborative Filtering, a representative recommendation system methodology, consists of two approaches: neighbor methods and latent factor models. Among these, the latent factor model using matrix factorization decomposes the user-item interaction matrix into two lower-dimensional rectangular matrices, predicting the item's rating through the product of these matrices. Due to the factor vectors inferred from rating patterns capturing user and item characteristics, this method is superior in scalability, accuracy, and flexibility compared to neighbor-based methods. However, it has a fundamental drawback: the need to reflect the diversity of preferences of different individuals for items with no ratings. This limitation leads to repetitive and inaccurate recommendations. The Adaptive Deep Latent Factor Model (ADLFM) was developed to address this issue. This model adaptively learns the preferences for each item by using the item description, which provides a detailed summary and explanation of the item. ADLFM takes in item description as input, calculates latent vectors of the user and item, and presents a method that can reflect personal diversity using an attention score. However, due to the requirement of a dataset that includes item descriptions, the domain that can apply ADLFM is limited, resulting in generalization limitations. This study proposes a Generalized Adaptive Deep Latent Factor Recommendation Model, G-ADLFRM, to improve the limitations of ADLFM. Firstly, we use item ID, commonly used in recommendation systems, as input instead of the item description. Additionally, we apply improved deep learning model structures such as Self-Attention, Multi-head Attention, and Multi-Conv1D. We conducted experiments on various datasets with input and model structure changes. The results showed that when only the input was changed, MAE increased slightly compared to ADLFM due to accompanying information loss, resulting in decreased recommendation performance. However, the average learning speed per epoch significantly improved as the amount of information to be processed decreased. When both the input and the model structure were changed, the best-performing Multi-Conv1d structure showed similar performance to ADLFM, sufficiently counteracting the information loss caused by the input change. We conclude that G-ADLFRM is a new, lightweight, and generalizable model that maintains the performance of the existing ADLFM while enabling fast learning and inference.

• Korean Journal of Remote Sensing
• /
• v.39 no.3
• /
• pp.363-370
• /
• 2023

Various earth observation satellites need to provide accurate and high-quality data after launch. To maintain and enhance the quality of satellite data, it is crucial to employ a cross-calibration process that accounts for differences in sensor characteristics, such as the spectral band adjustment factor (SBAF). In this study, we utilized Landsat-8 and Sentinel-2A satellite imagery collected from desert sites in Libya4, Algeria3, and Mauritania2 among pseudo-invariant calibration sites to calculate and apply SBAF, thereby compensating the uncertainties arising from variations in bandwidths. We quantitatively compared the reflectance differences based on the similarity of bandwidths, including Blue, Green, Red, and both the near-infrared (NIR) narrow, and NIR bands of Sentinel-2A. Following the application of SBAF, significant results with reflectance differences of approximately 1% or less were observed for all bands except NIR. In the case of the Sentinel-2A NIR band, it exhibited a significantly larger bandwidth difference compared to the NIR narrow band. However, after applying SBAF, the reflectance difference fell within the acceptable error range (5%) of 1-2%. It indicates that SBAF can be applied even when there is a substantial difference in the bandwidths of the two sensors, particularly in situations where satellite utilization is limited. Therefore, it was determined that SBAF could be applied even when the bandwidth difference between the two sensors is large in a situation where satellite utilization is limited. It is expected to be helpful in research utilizing the quality and continuity of satellite data.

• Journal of Internet Computing and Services
• /
• v.24 no.2
• /
• pp.19-26
• /
• 2023

Humans have adapted and evolved to natural light. However, as humans stay in indoor longer in modern times, the problem of biorhythm disturbance has been induced. To solve this problem, research is being conducted on lighting that reproduces the correlated color temperature(CCT) of natural light that varies from sunrise to sunset. In order to reproduce the CCT of natural light, multiple LED light sources with different CCTs are used to produce lighting, and then a control index DB is constructed by measuring and collecting the light characteristics of the combination of input currents for each light source in hundreds to thousands of steps, and then using it to control the lighting through the light characteristic matching method. The problem with this control method is that the more detailed the steps of the combination of input currents, the more time and economic costs are incurred. In this paper, an LED lighting control method that applies interpolation and combination calculation based on the minimum spectral power distribution information for each light source is proposed to reproduce the CCT of natural light. First, five minimum SPD information for each channel was measured and collected for the LED lighting, which consisted of light source channels with different CCTs and implemented input current control function of a 256-steps for each channel. Interpolation calculation was performed to generate SPD of 256 steps for each channel for the minimum SPD information, and SPD for all control combinations of LED lighting was generated through combination calculation of SPD for each channel. Illuminance and CCT were calculated through the generated SPD, a control index DB was constructed, and the CCT of natural light was reproduced through a matching technique. In the performance evaluation, the CCT for natural light was provided within the range of an average error rate of 0.18% while meeting the recommended indoor illumination standard.

• Korean Journal of Remote Sensing
• /
• v.39 no.2
• /
• pp.207-221
• /
• 2023

This study suggests deep neural network models for estimating air temperature with Level 1B (L1B) datasets of GEO-KOMPSAT-2A (GK-2A). The temperature at 1.5 m above the ground impact not only daily life but also weather warnings such as cold and heat waves. There are many studies to assume the air temperature from the land surface temperature (LST) retrieved from satellites because the air temperature has a strong relationship with the LST. However, an algorithm of the LST, Level 2 output of GK-2A, works only clear sky pixels. To overcome the cloud effects, we apply a deep neural network (DNN) model to assume the air temperature with L1B calibrated for radiometric and geometrics from raw satellite data and compare the model with a linear regression model between LST and air temperature. The root mean square errors (RMSE) of the air temperature for model outputs are used to evaluate the model. The number of 95 in-situ air temperature data was 2,496,634 and the ratio of datasets paired with LST and L1B show 42.1% and 98.4%. The training years are 2020 and 2021 and 2022 is used to validate. The DNN model is designed with an input layer taking 16 channels and four hidden fully connected layers to assume an air temperature. As a result of the model using 16 bands of L1B, the DNN with RMSE 2.22℃ showed great performance than the baseline model with RMSE 3.55℃ on clear sky conditions and the total RMSE including overcast samples was 3.33℃. It is suggested that the DNN is able to overcome cloud effects. However, it showed different characteristics in seasonal and hourly analysis and needed to append solar information as inputs to make a general DNN model because the summer and winter seasons showed a low coefficient of determinations with high standard deviations.