• Journal of the Korea Society of Computer and Information
• /
• v.27 no.2
• /
• pp.33-42
• /
• 2022

In this paper, we proposed a deep learning model based on multi-scale dilated convolution feature fusion for the segmentation of black ice region in road image to send black ice warning to drivers in real time. In the proposed multi-scale dilated convolution feature fusion network, different dilated ratio convolutions are connected in parallel in the encoder blocks, and different dilated ratios are used in different resolution feature maps, and multi-layer feature information are fused together. The multi-scale dilated convolution feature fusion improves the performance by diversifying and expending the receptive field of the network and by preserving detailed space information and enhancing the effectiveness of diated convolutions. The performance of the proposed network model was gradually improved with the increase of the number of dilated convolution branch. The mIoU value of the proposed method is 96.46%, which was higher than the existing networks such as U-Net, FCN, PSPNet, ENet, LinkNet. The parameter was 1,858K, which was 6 times smaller than the existing LinkNet model. From the experimental results of Jetson Nano, the FPS of the proposed method was 3.63, which can realize segmentation of black ice field in real time.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.34 no.6
• /
• pp.1779-1786
• /
• 2014

Artificial Intelligence-based techniques have been applied to problems where mathematical relations can not be presented due to complicatedness of the physical process. A representative example in hydraulics is the local scour around bridge piers. This study presents a GEP model for predicting the local scour around bridge piers. The model is trained by 64 laboratory data to build the regression equation, and the constructed model is verified against 33 laboratory data. Comparisons between the models with dimensional and normalized variables reveals that the GEP model with dimensional variables predicts better. The proposed model is now applied to two field datasets. It is found that the MAPE of the scour depths predicted by the GEP model increases compared with the predictions of local scours in laboratory scale. In addition, the model performance increases significantly when the model is trained by the field dataset rather than the laboratory dataset. The findings suggest that apart from the ANN model, GEP model is a sound and reliable model for predicting local scour depth.

• Journal of the Korean Solar Energy Society
• /
• v.27 no.4
• /
• pp.1-9
• /
• 2007

Sensitivity analysis of wind resource micrositing has been performed through the application case at the Antarctic King Sejong station with the most representative micrositing softwares: WAsP, WindSim and Meteodyn WT. The wind data obtained from two met-masts separated 625m were applied as a climatology input condition of micro-scale wind mapping. A tower shading effect on the met-mast installed 20m apart from the warehouse has been assessed by the CFD software Fluent and confirmed a negligible influence on wind speed measurement. Theoretically, micro-scale wind maps generated by the two met-data located within the same wind system and strongly correlated meteor-statistically should be identical if nothing influenced on wind prediction but orography. They, however, show discrepancies due to nonlinear effects induced by surrounding complex terrain. From the comparison of sensitivity analysis, Meteodyn WT employing 1-equation turbulence model showed 68% higher RMSE error of wind speed prediction than that of WindSim using the ${\kappa}-{\epsilon}$ turbulence model, while a linear-theoretical model WAsP showed 21% higher error. Consequently, the CFD model WindSim would predict wind field over complex terrain more reliable and less sensitive to climatology input data than other micrositing models. The auto-validation method proposed in this paper and the evaluation result of the micrositing softwares would be anticipated a good reference of wind resource assessments in complex terrain.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.6
• /
• pp.57-65
• /
• 2018

Conventional RF signal-based indoor localization techniques such as BLE or Wi-Fi based fingerprinting method show considerable localization errors even in small-scale indoor environments due to unstable received signal strength(RSS) of RF signals. Therefore, it is difficult to apply the existing RF-based fingerprinting techniques to large-scale indoor environments such as airports and department stores. In this paper, instead of RF signal we use the geomagnetic sensor signal for indoor localization, whose signal strength is more stable than RF RSS. Although similar geomagnetic field values exist in indoor space, an object movement would experience a unique sequence of the geomagnetic field signals as the movement continues. We use a deep neural network model called the recurrent neural network (RNN), which is effective in recognizing time-varying sequences of sensor data, to track the user's location and movement path. To evaluate the performance of the proposed geomagnetic field based indoor positioning system (IPS), we constructed a magnetic field map for a campus testbed of about $94m{\times}26$ dimension and trained RNN using various potential movement paths and their location data extracted from the magnetic field map. By adjusting various hyperparameters, we could achieve an average localization error of 1.20 meters in the testbed.

• Journal of Biosystems Engineering
• /
• v.28 no.3
• /
• pp.199-208
• /
• 2003

In order to solve the shortage of roughage supply for dairy farm in Korea, winter cereal forage production after harvesting of rice in the fallow paddy field is studied. This study consist of two parts. One is the model development of the mechanized production which was already reported at the preview paper. This is the 2nd parts of the study. Also, the mechanized production model for cereal forage production in winter was reported in the previous study. In this paper, coverage area and mechanized wrap silge production cost are analyzed and compared to the other available feeds in Korea. Results of the research are summarized as follows; The coverage area for the winter cereal wrap silage production system in Korea is estimated to be 33.7 ha in case of working with a tractor and a set of implements. If two or three tractors are available, the coverage area is estimated to be 68.0 and 101.3 ha, respectively. The break even point (BEP) of the farming size is analyzed as 10 ha and its production(operating) cost is estimated to be 317 to 443 won/TDN-kg at the BEP point. The cost is lowered to 182 won/TDN-kg at 100ha-working, and is much lower compared with prices of imported feeds of 360∼600 won/TDN-kg. Therefore, winter cereal wrap silage model is judged to be feasible and desirable for a large scale production of forage in winter fallow paddy field.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.8 no.2
• /
• pp.159-166
• /
• 2010

The block-scale groundwater flow system at Olkiluoto site in Finland was simulated. The heterogeneous and anisotropic hydraulic conductivity field for the domain was constructed from the discrete fracture network, which considered only the fractured zones identified in the deep boreholes installed in the study site. The groundwater flow model was calibrated by adjusting the recharge rate and the transmissivities of the fractured zones to fit the calculated hydraulic heads and into- and out-flow rates in the observation intervals of the boreholes with the observed ones. In the calibrated model, the calculated flow rates at some intervals were not in accordance with the observed ones although the calculated hydraulic heads fit well with the observed ones, which revealed that the number of the conduits for groundwater flow is insufficient in the conceptual model for groundwater flow modeling. Therefore, it was recommended that the potential local conduits such as background fractures should be added to the present conceptual model.

• Korean Journal of Health Education and Promotion
• /
• v.6 no.2
• /
• pp.23-39
• /
• 1989

There have been a lot of considerable. discussion and debate surrounding the management model in the health insurance management system and opinions regarding the management operating cost. It is a well known fact that there have always been dissenting opinions and debates surrounding the issue. The management operating cost varies according to the scale of the management organization and component members characteristics of the insurance carrier. Therefore, it is necessary to examine and compare the management operating cost to the simulated management models developed to cover those eligible for the health insurance scheme in this country. Since the management operating cost can vary according to the different models of management, four alternative management models have been established based on the critical evaluation of existing theories concerned, as well as on the basis of the survey results and simulation attempts. The first alternative model is the Unique Insurance Carrier Model(Ⅰ) ; desigened to cover all of the people with no classification of insurance qualifications and finances from the source of contribution of the insured, nationwide. The second is the Management Model of Large-scale District Insurance Carrier(Ⅱ) ; this means the Korean society would be divided into 21 large districts; each having its own insurance carrier that would cover the people in that particular district with no classification of insurance qualifications arid finances as in Model I. The third is the Management Model of Insurance Carrier Divided by Area and Classified with Occupation if Largescale (Ⅲ) ; to serve the self-employed in the 21 districts divided as in Model Ⅱ. It would serve the employees and their dependents by separate insurance carriers in large-scale similar to the area of the district-scale for the self-employed, so that the insurance qualifications and finances would be classified with each of the insurance carriers: The last is the Management Model of the Multi - insurance Carrier (Ⅳ) based on the Si. Gun. Gu area which will cover their own self- employed people in the area with more than 150 additional insurance carriers covering the employees and their dependents. The manpower necessary to provide services to all of the people according to the four models is calculated through simulation trials. It indicates that the Management Model of Large-scale District Insurance Carrier requires the most manpower among the four alternative models. The unit management operating costs per the insured individuals and covered persons are leveled with several intervals based on the insurance recipients. in their characteristics. The interval levels derived from the regression analysis reveal that the larger the scale of the insurance carriers is in the number of those insured and covered. the more the unit management operating cost decreases. significantly. Moreover. the result of the quadratic functional formula also shows the U-shape significantly. The management operating costs derived from the simulated calculation. on the basis of the average salary and related cost per staff- member of the Health Insurance Societies for Occupational Labours and Korean Medical Insurance Corporation for the Official Servants and Private School Teachers in 1987 fiscal year. show that the Model of Multi-insurance Carrier warrants the highest management operating cost. Meanwhile the least expensive management operating cost is the Management Model of Unique Insurance Carrier. Insurance Carrier Divided by Area and Classified with Occupation in Large-scale. and Large-scale District Insurance Carrier. in order. Therefore. it is feasible to select the Unique Insurance Carrier Model among the four alternatives from the viewpoint of the management operating cost and in the sense of the flexibility in promoting the productivity of manpower in the human services field. However. the choice of the management model for health insurance systems and its application should be examined further utilizing the operation research analysis for such areas as the administrative efficiency and factors related to computer cost etc.

• Journal of KIISE
• /
• v.41 no.11
• /
• pp.927-934
• /
• 2014

In this paper, we propose an effective method for recognizing daily human activities from a stream of three dimensional body poses, which can be obtained by using Kinect-like RGB-D sensors. The body pose data provided by Kinect SDK or OpenNI may suffer from both the view variance problem and the scale variance problem, since they are represented in the 3D Cartesian coordinate system, the origin of which is located on the center of Kinect. In order to resolve the problem and get the view-invariant and scale-invariant features, we transform the pose data into the spherical coordinate system of which the origin is placed on the center of the subject's hip, and then perform on them the scale normalization using the length of the subject's arm. In order to represent effectively complex internal structures of high-level daily activities, we utilize Hidden state Conditional Random Field (HCRF), which is one of probabilistic graphical models. Through various experiments using two different datasets, KAD-70 and CAD-60, we showed the high performance of our method and the implementation system.

• Wind and Structures
• /
• v.25 no.3
• /
• pp.233-259
• /
• 2017

Investigations on simulated near-surface atmospheric boundary layer (ABL) in an open-jet facility are carried out by conducting experimental tests on small-scale models of low-rise buildings. The objectives of the current study are: (1) to determine the optimal location of test buildings from the exit of the open-jet facility, and (2) to investigate the scale effect on the aerodynamic pressure characteristics. Based on the results, the newly built open-jet facility is well capable of producing mean wind speed and turbulence profiles representing open-terrain conditions. The results show that the proximity of the test model to the open-jet governs the length of the separation bubble as well as the peak roof pressures. However, test models placed at a horizontal distance of 2.5H (H is height of the wind field) from the exit of the open-jet, with a width that is half the width of the wind field and a length of 1H, have consistent mean and peak pressure coefficients when compared with available results from wind tunnel testing. In addition, testing models with as large as 16% blockage ratio is feasible within the open-jet facility. This reveals the importance of open-jet facilities as a robust tool to alleviate the scale restrictions involved in physical investigations of flow pattern around civil engineering structures. The results and findings of this study are useful for putting forward recommendations and guidelines for testing protocols at open-jet facilities, eventually helping the progress of enhanced standard provisions on the design of low-rise buildings for wind.

• Asia pacific journal of information systems
• /
• v.6 no.1
• /
• pp.241-267
• /
• 1996

The problems like exceeding estimated cost, late due-date, expensive maintenance, insuffiency of experts and low productivity are becoming emphasized related with the software development. In order to overcome these problems and to develop the highly qualified software within the limited resources, a project management tool is used. The purpose of this study is to develop a conceptual model of the project management system which can plan, analyze and control the software development projects effectively. The model is constructed with plan and track management system, cost management system and operation management system for the efficient project management. It is named by LSD-PMS : Large-scale Software Development Project Management System. We reviewed 5 cases of project management. LSD-PMS is evaluated and reviewed by the project managers in the field. It is proved that LSD-PMS is a tool which can help project managers develop software successfully given budget and time schedule. In the future, this system should be further developed as an integrated model with system implementation tools such as CASE products.