• Journal of the Korean Association of Geographic Information Studies
• /
• v.16 no.3
• /
• pp.103-114
• /
• 2013

In this study, we established the ortho mosaic imagery on the Korean Peninsula using KOMPSAT-2 images and conducted an accuracy assessment. Rational Polynomial Coefficient(RPC) modeling results were mostly less than 2 pixels except for mountainous regions which was difficult to select a Ground Control Point(GCP). Digital Elevation Model(DEM) which was made using the digital topographic map on the scale of 1:5,000 was used for generating an ortho image. In the case of inaccessible area, the Shuttle Radar Topography Mission(SRTM) DEM was used. Meanwhile, the ortho mosaic image of the Korean Peninsula was produced by each ortho image aggregation and color adjustment. An accuracy analysis for the mosaic image was conducted about a 1m color fusion image. In order to verify a geolocation accuracy, 813 check points which were acquired by field survey in South Korea were used. We found that the maximum error was not to exceed 5m(Root Mean Square Error : RMSE). On the other hand, in the case of inaccessible area, the extracted check points from a reference image were used for accuracy analysis. Approximately 69% of the image has a positional accuracy of less than 3m(RMSE). We found that the seam-line accuracy among neighboring image was very high through visual inspection. However, there were a discrepancy with 1 to 2 pixels at some mountainous regions.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.16 no.3
• /
• pp.246-260
• /
• 2014

A 90-m horizontal-resolution numerical model was configured to study the micrometeorological features of local winds in the valley of Gwangneung KoFlux (Korea Flux network) Sites (GDK: Gwangneung Deciduous forest site in Korea, GCK: Gwangneung Coniferous forest site in Korea) during summer days. The U. S. Geological Survey (USGS) Shuttle Radar Topography Mission (SRTM) data were employed for high-resolution model terrain height. Model performance was evaluated by comparing observed and simulated near-surface temperature and winds. Detailed qualitative analysis of the model-simulated wind field was carried out for two selected cases which are a clear day (Case I) and a cloudy day (Case II). Observed winds exhibited that GDK and GCK, as well as Case I and Case II, had differences in timing, duration and strength of daytime and nighttime wind direction and speeds. The model simulation results strongly supported the existence of the drainage flow in the valley of the KoFlux tower sites. Overall, the simulated model fields realistically presented the diurnal cycle of local winds in and around the valley, including the morning drainage-upslope transition and the evening reversal of upslope wind. Also, they indicated the complexity of local winds interactions by presenting that daytime westerly winds in the valley were not always pure mountain winds and were often coupled with larger-scale wind systems, such as synoptic-scale winds or mesoscale sea breezes blowing from the west coast of the peninsula.

• Korean Journal of Remote Sensing
• /
• v.35 no.6_2
• /
• pp.1037-1052
• /
• 2019

We had developed in preceding study a classification model for the Korean pine and Larch with an accuracy of 98 percent using Hyperion and Sentinel-2 satellite images, texture information, and geometric information as the first step for tree species mapping in the inaccessible North Korea. Considering a share of major tree species in North Korea, the classification model needs to be expanded as it has a large share of Oak(29.5%), Pine (12.7%), Fir (8.2%), and as well as Larch (17.5%) and Korean pine (5.8%). In order to classify 5 major tree species, national forest type map of South Korea was used to build 11,039 training and 2,330 validation data. Sentinel-2 data was used to derive spectral information, and PlanetScope data was used to generate texture information. Geometric information was built from SRTM DEM data. As a machine learning algorithm, Random forest was used. As a result, the overall accuracy of classification was 80% with 0.80 kappa statistics. Based on the training data and the classification model constructed through this study, we will extend the application to Mt. Baekdu and North and South Goseong areas to confirm the applicability of tree species classification on the Korean Peninsula.

• Korean Journal of Remote Sensing
• /
• v.39 no.5_1
• /
• pp.495-505
• /
• 2023

The Musan mine, situated in Musan County, Hamgyong Province, North Korea, stands as a prominent open-pit iron mine on the Korean Peninsula. This study focuses on estimating the mining and dumping activities within the Musan mine area by analyzing digital elevation model (DEM) changes. To calculate the long-term volume changes in the Musan mine, we digitized and converted the 1:200,000-scale third topographic map of the Joseon published in 1918 and compared with interferometric synthetic aperture radar (InSAR) DEMs, including Shuttle Radar Topography Mission DEM (2000) and Copernicus DEM (2011-2015). The findings reveal that over a century, Musan mine yielded around 1.37 billion tons of iron ore, while approximately 1.06 billion tons of waste rock were dumped. This study is particularly significant as it utilizes a historical topographic map predating the full-scale development of Musan mine to estimate a century's mining production and waste rock deposition. It is expected that this research provides valuable insights for future investigation of surface change of North Korea where the acquisition of in situ data remains challenging.

• 한국지형공간정보학회:학술대회논문집
• /
• 2005.08a
• /
• pp.93-99
• /
• 2005

This paper summarizes a preliminary geoid computation for western Mongolia, utilizing the airborne data collected fall 2004, as part of the NGA-DNSC-ALAGaC-MonMap cooperative airborne gravity project. A gravimetric geoid has been computed using the airborne gravity data, SRTM terrain models and GRACE/EGM global fields. The gravimetric geoid has subsequently been fitted to GPS-leveling data across Western Mongolia, as well as for a special Ulaanbaatar city geoid model.

• Journal of Korean Society for Geospatial Information Science
• /
• v.22 no.4
• /
• pp.13-19
• /
• 2014

To develop a wind farm in a mountainous terrain like Korea, it is generally more advantageous to install wind turbines along a mountain ridge where has relatively better wind resource because that is open in all directions and free from shielding by the surrounding topography. In this study, the SRTM (Shuttle Radar Topography Mission) v4.1 3 arc-second resolution digital elevation database and the geomorphometric characterization software LandSerf v2.3 are used to extract ridge lines for assessing a wind farm siting in mountainous terrain. The effectiveness of wind farm siting along a ridge line is confirmed that the most of wind turbines in the Gangwon, Taegisan, and Maebongsan wind farms in Korea's mountainous terrain are placed along the primary and secondary ridge lines where wind resource is relatively outstanding.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.26 no.5
• /
• pp.519-527
• /
• 2008

The basic elements in precise geoid determination are the gravity and topographic data with reliable quality and distribution. In this study, the effect of the gravity and topographic data on the precision of the geoid are analyzed through simulations in which the quality and distribution of the data are artificially controlled. It was found that the distribution of the topographic data has more effect on the precision of geoid than the quality of the it. This leads to the conclusion that the SRTM (Shuttle Radar Topography Mission) DTM (Digital Terrain Model) with resolution of 90m is qualified as a topographic data in geoid determination. In the experiments with gravity data, on the other hand, the aliasing effect caused by the low data density caused large errors in geoid. It was found that the more gravity data especially in north-eastern mountainous area is needed for precise geoid determination in Korea.

• Korean Journal of Remote Sensing
• /
• v.33 no.6_3
• /
• pp.1233-1242
• /
• 2017

In this study, we tried to analyze deforestation due to forest fire by using KOMPSAT satellite imagery. For deforestation analysis, unsupervised change detection algorithm is applied to multitemporal images. Through ITPCA (ITerative Principal Component Analysis) of NDVI (Normalized Difference Vegetation Index) generated from multitemporal satellite images before and after forest fire, changed areas due to deforestation are extracted. In addition, a post-processing method using SRTM (Shuttle Radar Topographic Mission) data is involved in order to minimize the error of change detection. As a result of the experiment using KOMPSAT-2 and 3 images, it was confirmed that changed areas due to deforestation can be efficiently extracted.

• Journal of Cadastre & Land InformatiX
• /
• v.47 no.2
• /
• pp.39-47
• /
• 2017

The SAR payload of the KOMPSAT-5 is equipped with an X-band (9.66GHz) microwave-based sensor. Especially, since it has a fixed antenna that can be electronically steered with respect to the azimuth and elevation planes, various applications are expected. This study evaluates the production performance and the accuracy of the DEM by producing DEM using the HR and UH mode images of KOMPSAT-5. To evaluate the production performance of the DEM, the sensitivity of DEM was assessed through a baseline analysis and $2{\pi}$ ambiguity; it was found to have good production performance. In addition, to evaluate the accuracy of the produced DEM, 30 check points were compared with SRTM data. As a result, STDEV ${\pm}15-20m$ accuracy was obtained. If the accuracy of the DEM is improved by adjusting the parameters of the filtering method or phase unwrapping method in the future, it will be possible to widely use the KOMPSAT-5 image for environmental and disaster monitoring.

• Proceedings of the KSRS Conference
• /
• 2007.10a
• /
• pp.449-451
• /
• 2007

In the commercially managed timber lands, the information such as height, age, stand density, canopy closure and leaf area index need to be collected periodically. Stand volume is the most fundamental information in the valuation of timber, however, stand age information is the primary element of forest inventory and these two are highly correlated. Conventional method of collecting stand age information by field surveys such as ring count method is accurate; however, it is expensive, labor-intensive and time consuming. In the present study it was aimed to collect stand age information using modem techniques in a commercially managed timberland situated in Tennessee, USA. The Landsat Thematic Mapper (TM), Enhanced Thematic Mapper (ETM+) of three different periods, Shuttle Radar Topography Mission (SRTM), National elevation dataset (NED) and field inventory data were used. Normalized difference vegetation index (NDVI) and Tasselled Cap (TC) transformation techniques were applied on the TM and ETM+ data. The regression analysis was carried out to identify the correlation between stand age and NDVI, TC. In the present study about 2,469 datasets were analyzed. The $R^{2}$ value for stand age estimation was 0.713. The NDVI, TC2 and TC3 were found to produce accurate timber age information.