• Korean Journal of Remote Sensing
• /
• v.34 no.6_3
• /
• pp.1415-1425
• /
• 2018

Kompsat-5 is the first Earth Observation Satellite which is equipped with an SAR in Korea. SAR images are generated by receiving signals reflected from an object by microwaves emitted from a SAR antenna. Because the wavelengths of microwaves are longer than the size of particles in the atmosphere, it can penetrate clouds and fog, and high-resolution images can be obtained without distinction between day and night. However, there is no color information in SAR images. To overcome these limitations of SAR images, colorization of SAR images using Cycle GAN, a deep learning model developed for domain translation, was conducted. Training of Cycle GAN is unstable due to the unsupervised learning based on unpaired dataset. Therefore, we proposed MS Cycle GAN applying multi-scale discriminator to solve the training instability of Cycle GAN and to improve the performance of colorization in this paper. To compare colorization performance of MS Cycle GAN and Cycle GAN, generated images by both models were compared qualitatively and quantitatively. Training Cycle GAN with multi-scale discriminator shows the losses of generators and discriminators are significantly reduced compared to the conventional Cycle GAN, and we identified that generated images by MS Cycle GAN are well-matched with the characteristics of regions such as leaves, rivers, and land.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.34 no.5
• /
• pp.1645-1653
• /
• 2014

In general, RGB, NIR, and thermal images are used for obtaining geospatial data. Such multiband images are collected via devices mounted on satellites or manned flights, but do not always meet users' expectations, due to issues associated with temporal resolution, costs, spatial resolution, and effects of clouds. We believe high-resolution, multiband images can be obtained at desired time points and intervals, by developing a payload suitable for a low-altitude, auto-piloted UAV. To achieve this, this study first established a low-cost, high-resolution multiband image collection system through developing a sensor and a payload, and collected geo-referencing data, as well as RGB, NIR and thermal images by using the system. We were able to obtain a 0.181m horizontal deviation and 0.203m vertical deviation, after analyzing the positional accuracy of points based on ortho mosaic images using the collected RGB images. Since this meets the required level of spatial accuracy that allows production of maps at a scale of 1:1,000~5,000 and also remote sensing over small areas, we successfully validated that the payload was highly utilizable.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.14 no.4
• /
• pp.51-56
• /
• 2014

Thesedays, there are many smart device applications developed, especially on the using various sensors included in the smart device. Smart devices have several sensors which are camera, GPS, mike, and communication module for collecting ubiquitous environment, and many applications are developed by using such sensors. In this paper, we developed the multimedia stream architecture and examined the smart device applications based on open source with front and back-end server clouds for developing the conceptual architecture. Also, we examined the back-end distributed servers, realtime multimedia stream transferring, multi-media store, and media relay for other server and smart devices. We test the examined architecture on the real target environment to collect the SIP initial setup time, media stream delay, and end-to-end play time. The test results show that there have good network operation environment to provide realtime multimedia services, and we need to improve the end-to-end play time by minimizing the initial setup time.

• The Bulletin of The Korean Astronomical Society
• /
• v.42 no.2
• /
• pp.43.3-43.3
• /
• 2017

NGC 6822 is a dwarf irregular galaxy in the Local Group and it is located in 500 kpc, further than the Large Magellanic Cloud and the Small Magellanic Cloud. Therefore, we can study star-forming processes by local condition in NGC 6822 instead of tidal force of the Galactic gravitational field. Hubble V is the brightest of several H II complexes in this galaxy. We observed Hubble V by using IGRINS attached on the 2.7 m telescope at the McDonald Observatory in Texas, US in May 2016. We performed a spectral mapping of $15^{{\prime}{\prime}}{\times} 7^{{\prime}{\prime}}$area on H and K bands, and detected emission lines of bright $Br{\gamma}\;{\lambda}2.1661{\mu}m$ and weak He I ${\lambda}2.0587{\mu}m$. Molecular hydrogen lines of 1-0S(1) ${\lambda}2.1218{\mu}m$, 2-1 S(1) ${\lambda}2.2477{\mu}m$, and 1-0 S(0) ${\lambda}2.2227{\mu}m$ was also detected. These emission lines show the structure of an ionized core and excited surface of clouds by far-ultraviolet photons, photodissociation region (PDR). We present three-dimensional maps of emission line distributions through multi slit scanning data and compare these results with the previous study. This presentation shows the physical structure of the star-forming regions and we discuss a PDR model and an evolution of Hubble V complex.

• The Bulletin of The Korean Astronomical Society
• /
• v.43 no.1
• /
• pp.45.1-45.1
• /
• 2018

Formation of filaments and subsequent dense cores in ISM is one of the essential questions to address in star formation. To investigate this scenario in detail, we recently started a molecular line survey namely 'Filaments, the Universal Nursery of Stars (FUNS)' toward nearby filamentary clouds in Gould Belt using TRAO 14m single dish telescope equipped with a 16 multi-beam array. In the present work, we report the first look results of kinematics of a low mass star forming region L1478 of California molecular cloud. This region is found to be consisting of long filaments with a hub-filament structure. We performed On-The-Fly mapping observations covering ~1.1 square degree area of this region using C18O(1-0) as a low density tracer and 0.13 square degree area using N2H+(1-0) as a high density tracer, respectively. CS (2-1) and SO (32-21) were also used simultaneously to map ~290 square arcminute area of this region. We identified 10 filaments applying Dendrogram technique to C18O data-cube and 13 dense cores using FellWalker and N2H+ data set. Basic physical properties of filaments such as mass, length, width, velocity field, and velocity dispersion are derived. It is found that filaments in L~1478 are velocity coherent and supercritical. Especially the filaments which are highly supercritical are found to have dense cores detected in N2H+. Non-thermal velocity dispersions derived from C18O and N2H+ suggest that most of the dense cores are subsonic or transonic while the surrounding filaments are transonic or supersonic. We concluded that filaments in L~1478 are gravitationally unstable which might collapse to form dense cores and stars. We also suggest that formation mechanism can be different in individual filament depending on its morphology and environment.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.9
• /
• pp.401-410
• /
• 2016

3D registration is a computer vision technique of aligning multi-view range images with respect to a reference coordinate system. Various 3D registration algorithms have been introduced in the past few decades. Iterative Closest Point (ICP) is one of the widely used 3D registration algorithms, where various modifications are available nowadays. In the ICP-based algorithms, the closest points are considered as the corresponding points. However, this assumption fails to find matching points accurately when the initial pose between point clouds is not sufficiently close. In this paper, we propose a new method to solve this problem using the 6D distance (3D color space and 3D Euclidean distances). Moreover, a color segmentation-based adaptive sampling technique is used to reduce the computational time and improve the registration accuracy. Several experiments are performed to evaluate the proposed method. Experimental results show that the proposed method yields better performance compared to the conventional methods.

• Journal of the Korea Computer Graphics Society
• /
• v.29 no.4
• /
• pp.17-24
• /
• 2023

In computer graphics and computer vision research and its applications, various representations of 3D objects, such as point clouds, voxels, or triangular meshes, are used depending on the purpose. The need for animating characters using these representations is also growing. In a typical animation pipeline called skeletal animation, "skinning weight painting" is required to determine how joints influence a vertex on the character's skin. In this paper, we introduce a neural network for automatically performing skinning weight painting for characters represented in various formats. We utilize signed distance fields (SDF) to handle different representations and employ graph neural networks and multi-layer perceptrons to predict the skinning weights for a given point.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.28 no.3D
• /
• pp.417-422
• /
• 2008

With standard orthorectification algorithms, one can produce unacceptable structure duplication in the orthophoto due to the double projection. Because of the abrupt height differences, such structure duplication is a frequently occurred phenomenon in the dense urban area which includes multi-history buildings. Therefore, occlusion area detection especially for the urban area is a critical issue in generation of true orthophoto. This paper deals with occlusion area detection with visible height based approach from aerial imagery and LiDAR. In order to accomplish this, a grid format DSM is produced from the point clouds of LiDAR. Next, visible height based algorithm is proposed to detect the occlusion area for each camera exposure station with DSM. Finally, generation of true orthophoto is presented with DSM and previously produced occlusion maps. The proposed algorithms are applied in the Purdue campus, Indiana, USA.

• Korean Journal of Remote Sensing
• /
• v.39 no.5_1
• /
• pp.507-519
• /
• 2023

Multi-temporal optical images have been utilized for time-series monitoring of croplands. However, the presence of clouds imposes limitations on image availability, often requiring a cloud removal procedure. This study assesses the applicability of various machine learning algorithms for effective cloud removal in optical imagery. We conducted comparative experiments by focusing on two key variables that significantly influence the predictive performance of machine learning algorithms: (1) land-cover types of training data and (2) temporal variability of land-cover types. Three machine learning algorithms, including Gaussian process regression (GPR), support vector machine (SVM), and random forest (RF), were employed for the experiments using simulated cloudy images in paddy fields of Gunsan. GPR and SVM exhibited superior prediction accuracy when the training data had the same land-cover types as the cloud region, and GPR showed the best stability with respect to sampling fluctuations. In addition, RF was the least affected by the land-cover types and temporal variations of training data. These results indicate that GPR is recommended when the land-cover type and spectral characteristics of the training data are the same as those of the cloud region. On the other hand, RF should be applied when it is difficult to obtain training data with the same land-cover types as the cloud region. Therefore, the land-cover types in cloud areas should be taken into account for extracting informative training data along with selecting the optimal machine learning algorithm.

• Korean Journal of Remote Sensing
• /
• v.37 no.6_1
• /
• pp.1631-1645
• /
• 2021

This research aims to provide the characteristics of the world's first active lidar sensor Atmospheric Laser Doppler Instrument (ALADIN) wind data and Geostationary Korea Multi Purpose Satellite 2A (GK2A) Atmospheric Motion Vector (AMV) data by comparing two wind data. As a result of comparing the data from September 2019 to August 1, 2020, The total number of collocated data for the AMV (using IR channel) and Mie channel ALADIN data is 177,681 which gives the Root Mean Square Error (RMSE) of 3.73 m/s and the correlation coefficient is 0.98. For a more detailed analysis, Comparison result considering altitude and latitude, the Normalized Root Mean Squared Error (NRMSE) is 0.2-0.3 at most latitude bands. However, the upper and middle layers in the lower latitudes and the lower layer in the southern hemispheric are larger than 0.4 at specific latitudes. These results are the same for the water vapor channel and the visible channel regardless of the season, and the channel-specific and seasonal characteristics do not appear prominently. Furthermore, as a result of analyzing the distribution of clouds in the latitude band with a large difference between the two wind data, Cirrus or cumulus clouds, which can lower the accuracy of height assignment of AMV, are distributed more than at other latitude bands. Accordingly, it is suggested that ALADIN wind data in the southern hemisphere and low latitude band, where the error of the AMV is large, can have a positive effect on the numerical forecast model.