• Journal of the Korean Institute of Intelligent Systems
• /
• 제21권6호
• /
• pp.749-754
• /
• 2011

In this study, we propose the design of optimized pRBFNNs-based face recognition system using two-dimensional Image and ASM algorithm. usually the existing 2 dimensional face recognition methods have the effects of the scale change of the image, position variation or the backgrounds of an image. In this paper, the face region information obtained from the detected face region is used for the compensation of these defects. In this paper, we use a CCD camera to obtain a picture frame directly. By using histogram equalization method, we can partially enhance the distorted image influenced by natural as well as artificial illumination. AdaBoost algorithm is used for the detection of face image between face and non-face image area. We can butt up personal profile by extracting the both face contour and shape using ASM(Active Shape Model) and then reduce dimension of image data using PCA. The proposed pRBFNNs consists of three functional modules such as the condition part, the conclusion part, and the inference part. In the condition part of fuzzy rules, input space is partitioned with Fuzzy C-Means clustering. In the conclusion part of rules, the connection weight of RBFNNs is represented as three kinds of polynomials such as constant, linear, and quadratic. The essential design parameters (including learning rate, momentum coefficient and fuzzification coefficient) of the networks are optimized by means of Differential Evolution. The proposed pRBFNNs are applied to real-time face image database and then demonstrated from viewpoint of the output performance and recognition rate.

• International Journal of Highway Engineering
• /
• 제11권4호
• /
• pp.59-68
• /
• 2009

Torsional behavior of concrete pavement slabs due to temperature and moisture effects is constrained by self weight and friction etc, and causes stress as the result. The stress due to humidity variation in the slab is difficult to calculate while that due to temperature variation can easily be calculated by a commercial structural analysis program. Thus, the slab behavior can be predicted more accurately if the humidity effect is converted to equivalent temperature and is used as an input of structural analysis. In this study, a concrete pavement slab was constructed and strains of the slab due to environmental loadings were measured for long-term period. Thermal strains were subtracted from the measured strains by using thermal expansion coefficient of the concrete measured in a laboratory. Shrinkage strains, the remained strains, was supposed as additional thermal strains to calculate imaginary temperature with equivalent effect of the shrinkage by dividing the shrinkage with the thermal expansion coefficient. An existing shrinkage model was modified by considering the self weight and friction to be used in another model which can convert differential shrinkage between top and bottom of the slab to equivalent temperature difference. Addition research efforts on tensile stress reduction according to steady increase in the compressive strains are warranted for more accurate stress calculation.

• Proceedings of the Korea Water Resources Association Conference
• /
• 한국수자원학회 2016년도 학술발표회
• /
• pp.260-260
• /
• 2016

Soil erosion is a very serious problem from agricultural as well as environmental point of view. Various computer models have been used to estimate soil erosion and assess erosion control practice. Universal Soil loss equation (USLE) is a popular model which has been used in many countries around the world. Erosivity (USLE R-factor) is one of the USLE input parameters to reflect impacts of rainfall in computing soil loss. Value of R factor depends upon Energy (E) and maximum rainfall intensity of specific period ($I30_{max}$) of that rainfall event and thus can be calculated using higher temporal resolution rainfall data such as 10 minute interval. But 10 minute interval rainfall data may not be available in every part of the world. In that case we can use hourly rainfall data to compute this R factor. Maximum 60 minute rainfall ($I60_{max}$) can be used instead of maximum 30 minute rainfall ($I30_{max}$) as suggested by USLE manual. But the value of Average annual R factor computed using hourly rainfall data needs some correction factor so that it can be used in USLE model. The objective of our study are to derive relation between averages annual R factor values using 10 minute interval and hourly rainfall data and to determine correction coefficient for R factor using hourly Rainfall data.75 weather stations of Korea were selected for our study. Ten minute interval rainfall data for these stations were obtained from Korea Meteorological Administration (KMA) and these data were changed to hourly rainfall data. R factor and $I60_{max}$ obtained from hourly rainfall data were compared with R factor and $I30_{max}$ obtained from 10 minute interval data. Linear relation between Average annual R factor obtained from 10 minute interval rainfall and from hourly data was derived with $R^2=0.69$. Correction coefficient was developed for the R factor calculated using hourly rainfall data.. Similarly, the relation was obtained between event wise $I30_{max}$ and $I60_{max}$ with higher $R^2$ value of 0.91. Thus $I30_{max}$ can be estimated from I60max with higher accuracy and thus the hourly rainfall data can be used to determine R factor more precisely by multiplying Energy of each rainfall event with this corrected $I60_{max}$.

• Journal of Broadcast Engineering
• /
• 제24권5호
• /
• pp.703-712
• /
• 2019

Single image Super-Resolution (SISR) aims to generate a visually pleasing high-resolution image from its degraded low-resolution measurement. In recent years, deep learning - based super - resolution methods have been actively researched and have shown more reliable and high performance. A typical method is WaveletSRNet, which restores high-resolution images through wavelet coefficient learning based on feature maps of images. However, there are two disadvantages in WaveletSRNet. One is a big processing time due to the complexity of the algorithm. The other is not to utilize feature maps efficiently when extracting input image's features. To improve this problems, we propose an efficient single image super resolution method, named RDB-WaveletSRNet. The proposed method uses the residual dense block to effectively extract low-resolution feature maps to improve single image super-resolution performance. We also adjust appropriated growth rates to solve complex computational problems. In addition, wavelet packet decomposition is used to obtain the wavelet coefficients according to the possibility of large scale ratio. In the experimental result on various images, we have proven that the proposed method has faster processing time and better image quality than the conventional methods. Experimental results have shown that the proposed method has better image quality by increasing 0.1813dB of PSNR and 1.17 times faster than the conventional method.

• Journal of Korea Water Resources Association
• /
• 제54권spc1호
• /
• pp.1037-1051
• /
• 2021

Biochemical oxygen demand (BOD) concentration, one of important water quality indicators, is treated as the measuring item for the ecological chapter in lakes and rivers. This investigation employed novel two-stage hybrid paradigm (i.e., wavelet-based gated recurrent unit, wavelet-based generalized regression neural networks, and wavelet-based random forests) to predict BOD concentration in the Dosan and Hwangji stations, South Korea. These models were assessed with the corresponding independent models (i.e., gated recurrent unit, generalized regression neural networks, and random forests). Diverse water quality and quantity indicators were implemented for developing independent and two-stage hybrid models based on several input combinations (i.e., Divisions 1-5). The addressed models were evaluated using three statistical indices including the root mean square error (RMSE), Nash-Sutcliffe efficiency (NSE), and correlation coefficient (CC). It can be found from results that the two-stage hybrid models cannot always enhance the predictive precision of independent models confidently. Results showed that the DWT-RF5 (RMSE = 0.108 mg/L) model provided more accurate prediction of BOD concentration compared to other optimal models in Dosan station, and the DWT-GRNN4 (RMSE = 0.132 mg/L) model was the best for predicting BOD concentration in Hwangji station, South Korea.

• Korean Journal of Remote Sensing
• /
• 제38권5_3호
• /
• pp.967-977
• /
• 2022

The occurrence and intensity of wildfires are increasing with climate change. Emissions from forest fire smoke are recognized as one of the major causes affecting air quality and the greenhouse effect. The use of satellite product and machine learning is essential for detection of forest fire smoke. Until now, research on forest fire smoke detection has had difficulties due to difficulties in cloud identification and vague standards of boundaries. The purpose of this study is to detect forest fire smoke using Level 1 and Level 2 data of Geostationary Environment Monitoring Spectrometer (GEMS), a Korean environmental satellite sensor, and machine learning. In March 2022, the forest fire in Gangwon-do was selected as a case. Smoke pixel classification modeling was performed by producing wildfire smoke label images and inputting GEMS Level 1 and Level 2 data to the random forest model. In the trained model, the importance of input variables is Aerosol Optical Depth (AOD), 380 nm and 340 nm radiance difference, Ultra-Violet Aerosol Index (UVAI), Visible Aerosol Index (VisAI), Single Scattering Albedo (SSA), formaldehyde (HCHO), nitrogen dioxide (NO₂), 380 nm radiance, and 340 nm radiance were shown in that order. In addition, in the estimation of the forest fire smoke probability (0 ≤ p ≤ 1) for 2,704 pixels, Mean Bias Error (MBE) is -0.002, Mean Absolute Error (MAE) is 0.026, Root Mean Square Error (RMSE) is 0.087, and Correlation Coefficient (CC) showed an accuracy of 0.981.

• Journal of Environmental Impact Assessment
• /
• 제32권4호
• /
• pp.203-213
• /
• 2023

The AERMOD model was the most used, accounting for 89.0%, based on the analysis of the environmental impact assessment reports published in the Environmental Impact Assessment Information Support System (EIASS) between 2021 and 2022. The mismatch of versions between AERMET and AERMOD was found to be 25.3%. There was the operational time discrepancy of 50.6% from industrial complexes, urban development projects between used in the model and applied in estimating pollutant emissions. The results of applying various versions of the AERMET and AERMOD models to both area sources and point sources in both simple and complex terrain in the Gunsan area showed similar values after AERMOD version 12 (15181). Emissions are assessed as 24-hour operation, and the predicted concentration in both simple and complex terrain when using the variable emission coefficient option that applies an 8-hour daytime operation in the model is lowered by 37.42% ~ 74.27% for area sources and by 32.06% ~ 54.45% for point sources. Therefore, to prevent the error in using the variable emission coefficient, it is required to clearly present the emission calculation process and provide a detailed explanation of the composition of modeling input data in the environmental impact assessment reports. Also, thorough reviews by special institutions are essential.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• 제28권6B호
• /
• pp.697-707
• /
• 2008

The grid-based KIneMatic wave STOrm Runoff Model (KIMSTORM) by Kim (1998) predicts the temporal variation and spatial distribution of overland flow, subsurface flow and stream flow in a watershed. The model programmed with C++ language on Unix operating system adopts single flowpath algorithm for water balance simulation of flow at each grid element. In this study, we attempted to improve the model by converting the code into FORTRAN 90 on MS Windows operating system and named as ModKIMSTORM. The improved functions are the addition of GAML (Green-Ampt & Mein-Larson) infiltration model, control of paddy runoff rate by flow depth and Manning's roughness coefficient, addition of baseflow layer, treatment of both spatial and point rainfall data, development of the pre- and post-processor, and development of automatic model evaluation function using five evaluation criteria (Pearson's coefficient of determination, Nash and Sutcliffe model efficiency, the deviation of runoff volume, relative error of the peak runoff rate, and absolute error of the time to peak runoff). The modified model adopts Shell Sort algorithm to enhance the computational performance. Input data formats are accepted as raster and MS Excel, and model outputs viz. soil moisture, discharge, flow depth and velocity are generated as BSQ, ASCII grid, binary grid and raster formats.

• Journal of the Korean Association of Geographic Information Studies
• /
• 제27권1호
• /
• pp.115-127
• /
• 2024

This research aimed to construct models with various structures based on the Transformer module and to perform land cover classification, thereby examining the applicability of the Transformer module. For the classification of land cover, the Unet model, which has a CNN structure, was selected as the base model, and a total of four deep learning models were constructed by combining both the encoder and decoder parts with the Transformer module. During the training process of the deep learning models, the training was repeated 10 times under the same conditions to evaluate the generalization performance. The evaluation of the classification accuracy of the deep learning models showed that the Model D, which utilized the Transformer module in both the encoder and decoder structures, achieved the highest overall accuracy with an average of approximately 89.4% and a Kappa coefficient average of about 73.2%. In terms of training time, models based on CNN were the most efficient. however, the use of Transformer-based models resulted in an average improvement of 0.5% in classification accuracy based on the Kappa coefficient. It is considered necessary to refine the model by considering various variables such as adjusting hyperparameters and image patch sizes during the integration process with CNN models. A common issue identified in all models during the land cover classification process was the difficulty in detecting small-scale objects. To improve this misclassification phenomenon, it is deemed necessary to explore the use of high-resolution input data and integrate multidimensional data that includes terrain and texture information.

• Journal of Sericultural and Entomological Science
• /
• 12호
• /
• pp.47-50
• /
• 1970

There are many factors affecting the development of sericultural industry in Korea. The setting of a rational pricing system for sericultural products is one of important activities of the Korean Government to improve the incentives to producers. The determination o： the prices for many years were based on the production costs including a certain level of profits. Some of cost items are in conflict both in cocoon producers and silk-reeling industries. Government officials have to evaluate these conflicting problems and estimate the consequences of their decisions. In this situation the final decision often became political decisions. This analysis is aimed at providing an alternative method of setting the prices of sericultural products. The criteria of the equilibrium employed in this analysis are based on economic principle which equilibrium condition is determined by the relationships between the marginal productivity of input factors and factor prices. In order to obtain the related information Cobb-Douglas'functions were fitted using KIST computer and data were obtained mostly from the Bank of Korea and the Ministry of Agriculture and Forestru, An important assumption is that "Opportunity Costs" of factors input in both cocoon production and silk-Peeling industries are same, The major finding s obtained are as followings. 1) The sum of coefficient of production elastisity in silk-reeling industries is greater than one. Silk-reeling industries are operating under the situation of increasing return to scale and it is, therefore, expected to develop the industries as the capital-intensive large scale. 2) The cocoon producing farmers are under the situations of the decreasing return to scale and it is expected to continue their cocoon farming as the labor-intensive small scale, assuming the present level of production technology. As the development of commercial farming, the resources input in cocoon production will be shifted to the production for higher profitable crops, 3) The price elastisity of production is higher in cocoon production than in silk-reeling industries. It is expected that the price changing effects on domestic production will be resulted from cocoon producers. 4) Based on analysis results of marginal productivities and the opportunity costs of resources, cocoon price for meeting equilibrium price condition is to be increased by 8-16 percent or standard price level of silk increased by 6-8 percent. There were the possibilities of over evaluation on opportunity cost of resources input in silk-reeling industries, or income transfered from the farmers to the industries. It is recommended that the prices for meeting equilibrium price conditions are to be determined by 72 percent for cocoon and 28 percent for silk-reeling costs, based on standard level of the exporting prices.