• Korean Journal of Remote Sensing
• /
• v.37 no.5_1
• /
• pp.1149-1161
• /
• 2021

With a trend of the utilization of computer vision for satellite images, cloud detection using deep learning also attracts attention recently. In this study, we conducted a U-Net cloud detection modeling using SPARCS (Spatial Procedures for Automated Removal of Cloud and Shadow) Cloud Dataset with the image data augmentation and carried out 10-fold cross-validation for an objective assessment of the model. Asthe result of the blind test for 1800 datasets with 512 by 512 pixels, relatively high performance with the accuracy of 0.821, the precision of 0.847, the recall of 0.821, the F1-score of 0.831, and the IoU (Intersection over Union) of 0.723. Although 14.5% of actual cloud shadows were misclassified as land, and 19.7% of actual clouds were misidentified as land, this can be overcome by increasing the quality and quantity of label datasets. Moreover, a state-of-the-art DeepLab V3+ model and the NAS (Neural Architecture Search) optimization technique can help the cloud detection for CAS500 (Compact Advanced Satellite 500) in South Korea.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_2
• /
• pp.1623-1631
• /
• 2022

Since oil spills can be a significant threat to the marine ecosystem, it is necessary to obtain information on the current contamination status quickly to minimize the damage. Satellite-based detection of marine oil spills has the advantage of spatiotemporal coverage because it can monitor a wide area compared to aircraft. Due to the recent development of computer vision and deep learning, marine oil spill detection can also be facilitated by deep learning. Unlike the existing studies based on Synthetic Aperture Radar (SAR) images, we conducted a deep learning modeling using PlanetScope optical satellite images. The blind test of the DeepLabV3+ model for oil spill detection showed the performance statistics with an accuracy of 0.885, a precision of 0.888, a recall of 0.886, an F1-score of 0.883, and a Mean Intersection over Union (mIOU) of 0.793.

• KIPS Transactions on Computer and Communication Systems
• /
• v.12 no.3
• /
• pp.119-126
• /
• 2023

Sarcopenia is not well known enough to be classified as a disease in 2021 in Korea, but it is recognized as a social problem in developed countries that have entered an aging society. The diagnosis of sarcopenia follows the international standard guidelines presented by the European Working Group for Sarcopenia in Older People (EWGSOP) and the d Asian Working Group for Sarcopenia (AWGS). Recently, it is recommended to evaluate muscle function by using physical performance evaluation, walking speed measurement, and standing test in addition to absolute muscle mass as a diagnostic method. As a representative method for measuring muscle mass, the body composition analysis method using DEXA has been formally implemented in clinical practice. In addition, various studies for measuring muscle mass using abdominal images of MRI or CT are being actively conducted. In this paper, we develop an AI image segmentation model based on abdominal images of CT with a relatively short imaging time for the diagnosis of sarcopenia and describe the multicenter validation. We developed an artificial intelligence model using U-Net that can automatically segment muscle, subcutaneous fat, and visceral fat by selecting the L3 region from the CT image. Also, to evaluate the performance of the model, internal verification was performed by calculating the intersection over union (IOU) of the partitioned area, and the results of external verification using data from other hospitals are shown. Based on the verification results, we tried to review and supplement the problems and solutions.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.12 no.4
• /
• pp.295-306
• /
• 2010

Contrary to an intact rock, the jointed rock mass shows strain-dependent deformation characteristics (elastic modulus and damping ratio). The maximum elastic modulus of a rock mass can be obtained from an elastic wave-based exploration in a small strain level and applied to seismic analyses. However, the assessment and application of the non-linear characteristics of rock masses in a small to medium strain level ($10^{-4}{\sim}0.5%$) have not been carried out yet. A non-linear dynamic analysis module is newly developed for FLAC3D to simulate strain-dependent shear modulus degradation and damping ratio amplification characteristics. The developed module is verified by analyzing the change of the Ricker wave propagation. Strain-dependent non-linear characteristics are obtained from disks of cored samples using a rock mass dynamic testing apparatus which can evaluate wave propagation characteristics in a jointed rock column. Using the experimental results and the developed non-linear dynamic module, seismic analyses are performed for the intersection of a shaft and an inclined tunnel. The numerical results show that vertical and horizontal displacements of non-linear analyses are larger than those of linear analyses. Also, non-linear analyses induce bigger bending compressive stresses acting on the lining. The bending compressive stress concentrates at the intersection part. The fundamental understanding of a strain-dependent jointed rock mass behavior is achieved in this study and the analytical procedure suggested can be effectively applied to field designs and analyses.

• Journal of Korean Society of Transportation
• /
• v.32 no.3
• /
• pp.198-206
• /
• 2014

According to the statistics of the National Police Agency, red light running (RLR) crashes represent a significant safety issue throughout Korea. This study deals with the RLR crashes occurred at signalized intersections in Cheongju. The objectives of this study are to comparatively analyze the characteristics of between RLR crashes and the Non-RLR crashes, and to find out factors using a Binary Logistic Regression(BLR) model. In pursuing the above, the study gives particular attentions to testing the differences between the above two groups with the data of 2,246 RLR/ 3,884 Non-RLR crashes (2007-2011). The main results are as follows. First, many RLR crashes were occurred in the nighttime and in going straight. Second, the difference between RLR and Non-RLR crashes were clearly defined by crash type, maneuver of vehicle before crash, age of driver (30s, 50s), alcohol use and accident pattern. Finally, a statistically significant model (Hosmer and Lemeshow test : 7.052, p-value : 0.531) was developed through the BLR model.

• Journal of the Korea Concrete Institute
• /
• v.21 no.1
• /
• pp.75-84
• /
• 2009

An analytical model for predicting the shear strength of prestressed concrete beams without shear reinforcement was developed, on the basis of the existing strain-based shear strength model. It was assumed that the compression zone of intact concrete in the cross-section primarily resisted the shear forces rather than the tension zone. The shear capacity of concrete was defined based on the material failure criteria of concrete. The shear capacity of the compression zone was evaluated along the inclined failure surface, considering the interaction with the compressive normal stress. Since the distribution of the normal stress varies with the flexural deformation of the beam, the shear capacity was defined as a function of the flexural deformation. The shear strength of a beam was determined at the intersection of the shear capacity curve and the shear demand curve. The result of the comparisons to existing test results showed that the proposed model accurately predicted the shear strength of the test specimens.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.18 no.6
• /
• pp.137-154
• /
• 2019

This study analysed the center control emergency vehicle preemption[EVP] test result on the 1.782 km section around Gangbuk Fire Station. The pros and cons between center control and site control EVP was compared through the review of existing research. The test site was selected based on the higher link speed for choosing low congested area and 4 to 6 lane road. EVP operates green extension under the estimated arrival time to each intersection. This study is about EVP system field application and its evaluation by analyzing EVP operation result with the emergency vehicle's trace, GPS data. The impact on the surrounding traffic was analysed in delay from the queue length survey. Analysis showed the decrease in averge travel time 41.81%, but the increase in delay of surrounding traffic slightly. It is expected that EVP can be applied to the expanded area by researching EVP compensation scheme.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.10 no.5
• /
• pp.14-22
• /
• 2011

This study is to estimate delay time model of signalized intersection by using travel data of Bus Information System. BIS, which applies the advanced information technology to an existing bus system, has been developing and operating in many cities. However, even though some useful traffic informations have been collected from BIS operation, utilization of real-time data to the traffic operation has not been promoted due to the inhomogeneity of modal speeds. Accordingly, in this study, a fundamental research is performed for traffic controls in urban areas and providing a traffic information throughout a methodology for estimating delay time using the data from BIS was developed. This delay time model setting bus travel time excluding service time of a bus stop as explanatory variables was constructed as a regression model, and the coefficient of determination of a linear regression model most highly appeared as 0.826. As a result of performing T-test with field survey values and model estimation values for verifying constructed models statistically, it was analyzed to be statistically significant in a confidence level of 95%.

• Archives of design research
• /
• v.15 no.3
• /
• pp.63-72
• /
• 2002

I made the three main estimated head about the facility of mechanical pencil which we always use in normal days. The head are these, the first is apparent usability by experiment, the second is the feeling when grasping, the third is the feeling when writing. After making the head, 1 selected the head which give the effect to the facility and make sure the interaction between heads by the method of actual test with 36 samples. After this test, 1 can find that there is no high interaction between the facility and the intuitively feeling by experiment. It's because the expectation which was made by experiment for the facility is higher than the real. but there is high interaction between the feeling of wearing and the feeling of writing. The mechanical pencil which has good facility is that the total length is long(169mm), the thickness is big($\Phi$8.5mm), the weight is little bit heavy(21.4g), the intersection is not circle, the surface where is hold by fingers has across grooves, and the body is glossy.

• Journal of Korea Multimedia Society
• /
• v.5 no.1
• /
• pp.18-27
• /
• 2002

A hierarchical representation scheme for planar curves, MBO-tree, is proposed in this paper, which provides natural approximation and efficient localization. MBO-tree is based on the Douglas-Peucker algorithm (iterative end-point fit algorithm), but approximation errors that are stored with corresponding points in MBO-tree nodes and are used for abstraction measures are adjusted by force to eliminate unnatural approximation. The error adjusting is just making the approximation error of a node in a MBO-tree to be less than or equal to that of its parent. In point of localization, the bounding area of a curve is represented with a minimum bounding octangle (MBO), which can enclose the curve more compactly compared with those of other hierarchical schemes, such as the strip tree, the arc tree and the HAL tree. The MBO satisfies the hierarchical inclusion property that is useful for hierarchical geometrical operations, such as the point-inclusion test and the polygon intersection test. Through several experiments, we found that the proposed scheme was able to approximate more naturally and to localize more effectively.