• Korean Journal of Remote Sensing
• /
• v.36 no.3
• /
• pp.487-501
• /
• 2020

As a viable option for retrieval of LST (Land Surface Temperature), this paper presents a DNN (Deep Neural Network) based approach using 148 Landsat 8 images for South Korea. Because the brightness temperature and emissivity for the band 10 (approx. 11-㎛ wavelength) of Landsat 8 are derived by combining physics-based equations and empirical coefficients, they include uncertainties according to regional conditions such as meteorology, climate, topography, and vegetation. To overcome this, we used several land surface variables such as NDVI (Normalized Difference Vegetation Index), land cover types, topographic factors (elevation, slope, aspect, and ruggedness) as well as the T₀ calculated from the brightness temperature and emissivity. We optimized four seasonal DNN models using the input variables and in-situ observations from ASOS (Automated Synoptic Observing System) to retrieve the LST, which is an advanced approach when compared with the existing method of the bias correction using a linear equation. The validation statistics from the 1,728 matchups during 2013-2019 showed a good performance of the CC=0.910~0.917 and RMSE=3.245~3.365℃, especially for spring and fall. Also, our DNN models produced a stable LST for all types of land cover. A future work using big data from Landsat 5/7/8 with additional land surface variables will be necessary for a more reliable retrieval of LST for high-resolution satellite images.

• Korean Journal of Remote Sensing
• /
• v.38 no.1
• /
• pp.103-110
• /
• 2022

Since satellite images generally include clouds in the atmosphere, it is essential to detect or mask clouds before satellite image processing. Clouds were detected using physical characteristics of clouds in previous research. Cloud detection methods using deep learning techniques such as CNN or the modified U-Net in image segmentation field have been studied recently. Since image segmentation is the process of assigning a label to every pixel in an image, precise pixel-based dataset is required for cloud detection. Obtaining accurate training datasets is more important than a network configuration in image segmentation for cloud detection. Existing deep learning techniques used different training datasets. And test datasets were extracted from intra-dataset which were acquired by same sensor and procedure as training dataset. Different datasets make it difficult to determine which network shows a better overall performance. To verify the effectiveness of the cloud detection network such as Cloud-Net, two types of networks were trained using the cloud dataset from KOMPSAT-3 images provided by the AIHUB site and the L8-Cloud dataset from Landsat8 images which was publicly opened by a Cloud-Net author. Test data from intra-dataset of KOMPSAT-3 cloud dataset were used for validating the network. The simulation results show that the network trained with KOMPSAT-3 cloud dataset shows good performance on the network trained with L8-Cloud dataset. Because Landsat8 and KOMPSAT-3 satellite images have different GSDs, making it difficult to achieve good results from cross-sensor validation. The network could be superior for intra-dataset, but it could be inferior for cross-sensor data. It is necessary to study techniques that show good results in cross-senor validation dataset in the future.

• Korean Journal of Remote Sensing
• /
• v.39 no.5_3
• /
• pp.1031-1042
• /
• 2023

In this study, a deep learning model was developed to predict the yield of cabbage and radish, one of the five major supply and demand management vegetables, using satellite images of Landsat 8. To predict the yield of cabbage and radish in Gangwon-do from 2015 to 2020, satellite images from June to September, the growing period of cabbage and radish, were used. Normalized difference vegetation index, enhanced vegetation index, lead area index, and land surface temperature were employed in this study as input data for the yield model. Crop yields can be effectively predicted using satellite images because satellites collect continuous spatiotemporal data on the global environment. Based on the model developed previous study, a model designed for input data was proposed in this study. Using time series satellite images, convolutional neural network, a deep learning model, was used to predict crop yield. Landsat 8 provides images every 16 days, but it is difficult to acquire images especially in summer due to the influence of weather such as clouds. As a result, yield prediction was conducted by splitting June to July into one part and August to September into two. Yield prediction was performed using a machine learning approach and reference models , and modeling performance was compared. The model's performance and early predictability were assessed using year-by-year cross-validation and early prediction. The findings of this study could be applied as basic studies to predict the yield of field crops in Korea.

• Journal of Environmental Impact Assessment
• /
• v.10 no.1
• /
• pp.1-8
• /
• 2001

Detection of red tides by satellite remote sensing can be done either by detecting enhanced level of chlorophyll pigment or by detecting changes in the spectral composition of pixels. Using chlorophyll concentration, however, is not effective currently due to the facts: 1) Chlorophyll-a is a universal pigment of phytoplankton, and 2) no accurate algorithm for chlorophyll in case 2 water is available yet. Here, red band algorithm, classification and PCA (Principal Component Analysis) techniques were applied for detecting patches of Cochlodinium polykrikoides red tides which occurred in Korean waters in 1995. This dinoflagellate species appears dark red due to the characteristic pigments absorbing lights in the blue and green wavelength most effectively. In the satellite image, the brightness of red tide pixels in all the three visible bands were low making the detection difficult. Red band algorithm is not good for detecting the red tide because of reflectance of suspended sediments. For supervised classification, selecting training area was difficult, while unsupervised classification was not effective in delineating the patches from surrounding pixels. On the other hand, PCA gave a good qualitative discrimination on the distribution compared with actual observation.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.8 no.1
• /
• pp.13-22
• /
• 2005

This paper classifies the topography of the Saemangeum Tidal flats based on Landsat TM satellite images by unsupervised ISODATA method, and analysis of the spatiotemporal changes of the classified shapes. The sedimental topography represents various properties according to the Saemangeum Tidal Embankment progress. We well proceed this study of the sedimental changes and distributions. By specifying the topographic characteristics of inner sea areas respectively, the investigation on the case study area according to the changes of the tidal will be useful in the establishment of land reclamation plan and the land use of the reclaimed area. In addition, the estuary image can be divided into tidal flats and sea surfaces using the band 4, also the detailed topography using the band 5, respectively among Landsat TM 7 bands. This paper contributes to the efficient image processing of the spatiotemporal sedimental changes.

• Korean Journal of Remote Sensing
• /
• v.34 no.2_1
• /
• pp.167-178
• /
• 2018

Urban green areas or forest regions play an important role in lowering the air temperature of the surrounding areas. This cooling effect does not only affect inside of the green areas, but also extends into neighboring streets and buildings. In this study, the Land Surface Temperature (LST) are retrieved from the Landsat 8 satellite data for 8 clear days in Seoul, Korea from 2013 to 2015, and used for analyzing the cooling effect at an urban green region, Seonjeongneung, located in the southern part of Seoul. The LST distribution from the boundary of the Seonjeongneung presents that the cooling effect of the green areas was found to extend in many directions into the urban areas. The LST estimations of residential and commercial areas around the Seonjeongneung are also analyzed to assess how the green areas affect the type of land cover and the surroundings in the urban areas. Relatively lower LST for the residential areas from the Seonjeongneung boundary ranges from 100 to 250 m, resulting in an average cooling effect of $2.3^{\circ}C$. On the other hand, the LST distribution in the commercial areas shows that the effective distance of green areas are relatively low in the range of 0 to 200 m, which means the average cooling effect is approximately $0.3^{\circ}C$. This result shows that the cooling effect of the Seonjeongneung is clearly noticeable, particularly, the residential areas show greater cooling effect than commercial areas.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.10 no.11
• /
• pp.1205-1216
• /
• 2015

Since concentration of greenhouse gas increases continuously from human's fossil fuel use, urbanization, and cultivation, it is trend that climate change is appearing. In Addition, in 20th century, occurrence of disaster is accidental and huge, and damage level also increases gradually. Therefore, in order to preserve the territory and to protect people's life and property against new type disasters, disaster detection satellite (payloads) development is required urgently. In this paper, we conduct a research and development for the prompt preemptive action when occurred a disaster, in particularly, about the disaster observation optimized at Korea's geographical features for the irregular future disasters. For the payload design which is specialized detect disasters, we create a tech tree of satellite imagery applications based 10 disaster types, and analyze the satellite sensor technologies referred to Landsat-8, Worldview-3 and ALOS-2.

• Proceedings of the KSRS Conference
• /
• 2001.03a
• /
• pp.116-121
• /
• 2001

경기만 지역은 세계최대 규모의 갯벌이 조성되어 해양생태계에서 중요한 역할을 수행하는 자연의 보고이나, 강한 조류운동, 한강 유역으로부터의 토사이동, 그리고 계속되는 연안 개발등 지속적인 영향을 받고 있다. 본 연구에서는 이러한 경기만 지역의 지리적·환경적 요인에 기인한 갯벌지역의 지난 30년 동안 공간적 변화를 분석하고자 한다. 해안선·조간대 지형의 변화 특성은 1972년부터 1999년까지 약 5년 간격으로 촬영된 Landsat MSS 와 TM 영상들을 이용하여 분석하였다. MSS와 TM의 공통적인 파장대이며, 물과 조간대의 경계가 뚜렷한 근적외선 파장대를 이용하여 간조시 갯벌의 경계선을 추출하였다. 각 시기의 수면, 갯벌, 육지를 나타내는 수치지도가 제작된 후, 이들을 중첩함으로써 시기별 변화유형을 구분하였고, 변화유형을 다시 원인에 따라 인공적인 요인과 자연적인 요인으로 나누었다. 의미있는 변화 유형은 크게 8가지로 나타났으며, 변화유형과 변화요인을 연계하여 경기만 지역의 변화특성을 도출하였다.

• Korean Journal of Remote Sensing
• /
• v.36 no.4
• /
• pp.627-635
• /
• 2020

With the increasing severity of climate change, intense torrential rains are occurring more frequently globally. Flooding due to torrential rain not only causes substantial damage directly, but also via secondary events such as landslides. Therefore, accurate and prompt flood detection is required. Because it is difficult to directly access flooded areas, previous studies have largely used satellite images. Traditionally, water indices such asthe normalized difference water index (NDWI) and modified normalized difference water index (MNDWI) which are based on different optical bands acquired by satellites, are used to detect floods. In addition, as flooding likelihood is greatly influenced by the weather, synthetic aperture radar (SAR) images have also been used, because these are less influenced by weather conditions. In this study, we compared flood areas calculated from SAR images and water indices derived from Landsat-8 images, where the images were acquired at similar times. The flooded area was calculated from Landsat-8 and Sentinel-1 images taken between the end of May and August 2019 at Lijiazhou Island, China, which is located in the Changjiang (Yangtze) River basin and experiences annual floods. As a result, the flooded area calculated using the MNDWI was approximately 21% larger on average than that calculated using the NDWI. In a comparison of flood areas calculated using water indices and SAR intensity images, the flood areas calculated using SAR images tended to be smaller, regardless of the order in which the images were acquired. Because the images were acquired by the two satellites on different dates, we could not directly compare the accuracy of the water-index and SAR data. Nevertheless, this study demonstrates that floods can be detected using both optical and SAR satellite data.

• Korean Journal of Remote Sensing
• /
• v.8 no.2
• /
• pp.131-145
• /
• 1992

A low cost personal computer and image processing S/W were empolyed to derive Digtal Elevation Model(DEM) of tidal flat from multitemporal LANDSAT TM images, and to create three-dimensional(3D) perspective views of the tidel flat on Komso bay in west coasts of Korea. The method for generation of Digital Elevation Model(DEM) in tidal flat was considered by overlapping techniques of multitemporal LANDSAT TM images and interpolations. The boundary maps of tidal flat extracted from multitemporal images with different water high were digitally combined in x, y, z space with tide in formation and used as an inputcontour data to obtain an elevation model by interpolation using spline function. Elevation errors of less than $\pm$0.1m were achived using overlapping techniques and a spline interpolation approach, respectively. The derived DEM allows for the generation of a perspective grid and drape on the satellite image values to create a realistic terrain visualization model so that the tidal flat may be viewed from and desired direction. As the result of this study, we obtained elevation model of tidal flats which contribute to characterize of topography and monitoring of morphological evolution of tidal flats. Moreover, the modal generated here can be used for simulation of innudation according to tide and support other studies as a supplementary data set.