Most Downloaded

  • December 31, 2022

    11 5

    Deep Learning-based Forest Fire Classification Evaluation for Application of CAS500-4

    차성은 1)·원명수 2)·장근창 3)·김경민 4)·김원국 5)·백승일 6)·임중빈 3)†

    Korean Journal of Remote Sensing 2022; 38(6): 1273-1283

    https://doi.org/10.7780/kjrs.2022.38.6.1.22

    Abstract
    최근 기후변화로 인해 중대형 산불이 빈번하게 발생하여 매년 인명 및 재산피해로 이어지고 있다. 원격 탐사를 활용한 산불 피해지 모니터링 기법은 신속한 정보와 대규모 피해지의 객관적인 결과를 취득할 수 있다. 본 연구에서는 산불 피해지를 분류하기 위해 Sentinel-2의 분광대역, 정규식생지수(normalized difference vegetation index, NDVI), 정규수역지수(normalized difference water index, NDWI)를 활용하여 2022년 3월 발생 한 강릉·동해 산불 피해지를 대상으로 U-net 기반 convolutional neural networks (CNNs) 딥러닝 모형을 모의하 였다. 산불 피해지 분류 결과 강릉·동해 산불 피해지의 경우 97.3% (f1=0.486, IoU= 0.946)로 분류 정확도가 높았 으나, 과적합(overfitting)의 가능성을 배제하기 어려워 울진·삼척 지역으로 동일한 모형을 적용하였다. 그 결과, 국립산림과학원에서 보고한 산불 피해 면적과의 중첩도가 74.4%로 확인되어 모형의 불확도를 고려하더라도 높은 수준의 정확도를 확인하였다. 본 연구는 농림위성과 유사한 분광대역을 선택적으로 사용하였으며, Sentinel-2 영상을 활용한 산불 피해지 분류가 정량적으로 가능함을 시사한다.
  • December 31, 2022

    13 5

    Observation Test of Field Surface Reflectance Using Vertical Rotating Goniometer on Tarp Surface and Grass

    문현동 1),2)·조은이1),2)·김현기3)·조유나3)·김보경4)·안호용5)·류재현6)·조재일 7),8)†

    Korean Journal of Remote Sensing 2022; 38(6): 1207-1217

    https://doi.org/10.7780/kjrs.2022.38.6.1.17

    Abstract
    농업의 새로운 패러다임인 디지털 농업에서는 원격탐사 기법을 이용하여 작물 생육을 지속으로 감시하 며 상태를 신속히 정보화하고 있다. 대표적으로 작물의 생육·생리적 변화에 대한 선택적인 파장 반사도 변화 를 기반으로 한 식생지수가 주로 사용되어진다. 하지만 식생 표면의 분광 반사도는 이방성을 갖기 때문에 태양 위치와 지면 관측 방향에 따라 변할 수 있어 식생지수 값이 작물의 상태를 나타내지 못하고 왜곡될 수 있다. 본 연구에서는 야외용 측각기를 제작하여 타프(tarp)와 잔디 식생에서 시험 운영하였다. 램버시안 표면과 유사한 성질의 타프에서는 Blue, Green, Red, 근적외선 파장에 대해 대체적으로 타프의 속성 반사도와 유사하게 측정 되었다. 하지만, 흐린 날은 센서 천정각이 커질수록 반사도가 과대 측정되는 경향을 보였다. 잔디에서 주요 식 생지수의 상대 차잇값을 보았을 때, 태양과 센서 천정각에 대해 visible atmospherically resistant index (VARI)와 vegetation index green (VIgreen), simple ratio (SR), normalized difference vegetation index (NDVI), enhanced vegetation index (EVI) 순으로 민감하였다. 측각기를 통해서 직접 관측된 자료들은 원격탐사 기반의 반사도 기반 식생지 수를 보다 정확하게 산출하는데 기여할 수 있을 것이다.
  • December 31, 2022

    9 5

    Post-processing Method of Point Cloud Extracted Based on Image Matching for Unmanned Aerial Vehicle Image

    이수암 1)†·김한결 2)·김태정 3)

    Korean Journal of Remote Sensing 2022; 38(6): 1025-1034

    https://doi.org/10.7780/kjrs.2022.38.6.1.4

    Abstract
    본 논문에서는 건물의 포인트 클라우드를 추출할 때 발생하는 홀 영역의 보간을 통한 후처리 방안을 제 안한다. 스테레오 영상 데이터에서 영상 매칭을 수행할 경우 차폐 및 건물 벽면 등의 영향으로 홀이 발생한다. 이런 영역은 추후 포인트 클라우드를 기반으로 하는 부가 산출물의 생성에 장애 요인이 될 수 있으므로, 이에 대한 효과적인 처리 기법의 적용이 필요하다. 먼저 영상 매칭을 적용하여 생성된 시차맵을 기반으로 초기 포인 트 클라우드를 추출한다. 포인트 클라우드를 격자화 시키면 차폐영역 및 건물 벽면의 영향으로 발생하는 홀 영 역을 확인할 수 있다. 홀 영역에 삼각망을 생성하고 삼각망 내부 값을 영역의 최소값으로 처리하는 과정을 반 복하는 것으로 건물 주변의 지표면과 건물 간에 어색함 없는 보간의 수행이 가능하다. 격자화 된 데이터에서 보간 된 영역에 해당하는 위치정보를 포인트로 추가하여 새로운 포인트 클라우드를 생성한다. 보간과정 중 불 필요한 점의 추가를 최소화하기 위해 초기 포인트 클라우드 영역에서 벗어나는 영역으로 보간 된 데이터는 처 리하지 않았으며, 보간 된 포인트 클라우드에 적용되는 RGB 밝기값은 매칭에 사용된 스테레오 영상 중 촬영 중심과 해당 픽셀이 가장 근접한 영상으로 설정하여 처리하였다. 실험 결과 제안 기법을 통해 대상영역의 포인 트 클라우드 생성 후 발생하는 음영 영역이 효과적으로 처리되는 것을 확인할 수 있었다.
  • December 31, 2022

    16 5

    Change Detection Using Deep Learning Based Semantic Segmentation for Nuclear Activity Detection and Monitoring

    송아람 1)·이창희2)·이진민3)·한유경 4)†

    Korean Journal of Remote Sensing 2022; 38(6): 991-1005

    https://doi.org/10.7780/kjrs.2022.38.6.1.1

    Abstract
    위성 영상은 핵 활동 탐지와 검증을 위한 효율적인 보조자료로 핵시설과 같이 접근이 어렵고 정보가 제 한된 지역에 매우 유용하다. 특히 장비의 이동 또는 시설물의 변화와 같이 핵실험을 준비하는 과정은 시계열 분석을 통해 충분히 식별 가능하다. 본 연구에서는 핵 활동과 관련된 주요 객체의 변화를 탐지하기 위하여, 다 시기 영상의 의미론적 분할 결과의 차이를 이용하였다. AIHub에서 제공하는 KOMPSAT 3/3A 영상으로 구성 된 객체 판독 데이터셋에서 건물, 도로, 소형 객체의 정보를 추출하여 학습하였으며, U-Net, PSPNet, Attention U-Net에 대하여 주요 파라미터를 변경하며 대상 객체 추출에 적합한 의미론적 분할 모델을 분석하였다. 의미 론적 분할 결과의 차영상으로 생성된 결과에 객체 정보를 포함하여 최종 변화 탐지를 수행하였으며, 제안 기법 을 임의의 변화를 포함한 시뮬레이션 영상에 적용한 결과, 변화 객체를 효과적으로 추출할 수 있었다. 본 연구 에서 제시된 변화 탐지 기법을 적용하기 위해서는, 의미론적 분할의 정확도가 우선적으로 확보되어야 하는 제 약이 있으나, 추후 실험 대상 지역에 대한 학습데이터셋이 증가할 수록 적용 가능한 분석 범위가 증가할 것으 로 기대된다.
  • ArticleOctober 31, 2022

    12 5

    Monitoring the Ecological Drought Condition ofVegetation during Meteorological Drought Using Remote Sensing Data

    Jeongeun Won 1)·Haeun Jung2)·Shinuk Kang3)·Sangdan Kim 4)†

    Korean Journal of Remote Sensing 2022; 38(5): 887-899

    https://doi.org/10.7780/kjrs.2022.38.5.3.7

    Abstract
    Drought caused by meteorological factors negatively affects vegetation in terrestrial ecosystems. In this study, the state in which meteorological drought affects vegetation was defined as the ecological drought of vegetation, and the ecological drought condition index of vegetation (EDCI- veg) was proposed to quantitatively monitor the degree of impact. EDCI-veg is derived from a copula- based bi-variate joint probability model between vegetation and meteorological drought information, and can be expressed numerically how affected the current vegetation condition was by the drought when the drought occurred. Comparing past meteorological drought events with their corresponding vegetation condition, the proposed index was examined, and it was confirmed that EDCI-veg could properly monitor the ecological drought of vegetation. In addition, it was possible to spatially identify ecological drought conditions by creating a high-resolution drought map using remote sensing data.
  • EditorialOctober 31, 2022

    18 5

    Remote Sensing and GIS for Earth & Environmental Disasters: The Current and Future in Monitoring, Assessment, and Management 2

    Minjune Yang 1)·Jae-Jin Kim 2)·Jong-Sik Ryu 3)· Kyung-soo Han 4)·Jinsoo Kim 5)†

    Korean Journal of Remote Sensing 2022; 38(5): 811-818

    https://doi.org/10.7780/kjrs.2022.38.5.3.1

    Abstract
    Recently, the number of natural and environmental disasters is rapidly increasing due to extreme weather caused by climate change, and the scale of economic losses and damage to human life is increasing accordingly. In addition, with urbanization and industrialization, the characteristics and scale of extreme weather appearance are becoming more complex and large in different ways from the past, and need for remote sensing and artificial intelligence technology for responding and managing global environmental disasters. This special issue investigates environmental disaster observation and management research using remote sensing and artificial intelligence technology, and introduces the results of disaster-related studies such as drought, flood, air pollution, and marine pollution, etc. in South Korea performed by the i-SEED (School of Integrated Science for Sustainable Earth and Environmental Disaster at Pukyong National University). In this special issue, we expect that the results can contribute to the development of monitoring and management technologies that may prevent environmental disasters and reduce damage in advance.
  • ArticleOctober 31, 2022

    5 5

    Exploitation of Dual-polarimetric Index of Sentinel-1 SAR Data inVessel Detection Utilizing Machine Learning

    Juyoung Song1)·Duk-jin Kim 2)†·Junwoo Kim3)·Chenglei Li1)

    Korean Journal of Remote Sensing 2022; 38(5): 737-746

    https://doi.org/10.7780/kjrs.2022.38.5.2.7

    Abstract
    Utilizing weather independent SAR images along with machine learning based object detector is effective in robust vessel monitoring. While conventional SAR images often applied amplitude data from Single Look Complex, exploitation of polarimetric parameters acquired from multiple polarimetric SAR images was yet to be implemented to vessel detection utilizing machine learning. Hence, this study used four polarimetric parameters (H, p1, DoP, DPRVI) retrieved from eigen-decomposition and two backscattering coefficients (γ0, VV, γ0, VH) from radiometric calibration; six bands in total were respectively exploited from 52 Sentinel-1 SAR images, accompanied by vessel training data extracted from AIS information which corresponds to acquisition time span of the SAR image. Evaluating different cases of combination, the use of polarimetric indexes along with amplitude values derived enhanced vessel detection performances than that of utilizing amplitude values exclusively.
  • ArticleOctober 31, 2022

    14 5

    Assessment of Stand-alone Utilization of Sentinel-1 SAR for High Resolution Soil Moisture Retrieval Using Machine Learning

    Jaehwan Jeong 1) · Seongkeun Cho 2) · Hyunho Jeon 3) · Seulchan Lee 2) · Minha Choi 4)†

    Korean Journal of Remote Sensing 2022; 38(5): 571-585

    https://doi.org/10.7780/kjrs.2022.38.5.1.11

    Abstract
    As the threat of natural disasters such as droughts, floods, forest fires, and landslides increases due to climate change, social demand for high-resolution soil moisture retrieval, such as Synthetic Aperture Radar (SAR), is also increasing. However, the domestic environment has a high proportion of mountainous topography, making it challenging to retrieve soil moisture from SAR data. This study evaluated the usability of Sentinel-1 SAR, which is applied with the Artificial Neural Network (ANN) technique, to retrieve soil moisture. It was confirmed that the backscattering coefficient obtained from Sentinel-1 significantly correlated with soil moisture behavior, and the possibility of stand-alone use to correct vegetation effects without using auxiliary data observed from other satellites or observatories. However, there was a large difference in the characteristics of each site and topographic group. In particular, when the model learned on the mountain and at flat land cross-applied, the soil moisture could not be properly simulated. In addition, when the number of learning points was increased to solve this problem, the soil moisture retrieval model was smoothed. As a result, the overall correlation coefficient of all sites improved, but errors at individual sites gradually increased. Therefore, systematic research must be conducted in order to widely apply high-resolution SAR soil moisture data. It is expected that it can be effectively used in various fields if the scope of learning sites and application targets are specifically limited.
  • ArticleOctober 31, 2022

    6 5
    Abstract
    This study is for the software implementation to generate atmospheric and surface reflectance products from RapidEye satellite imagery. The software is an extension based on Orfeo Toolbox (OTB) and an open-source remote sensing software including calibration modules which use an absolute atmospheric correction algorithm. In order to verify the performance of the program, the accuracy of the product was validated by a test image on the Radiometric Calibration Network (RadCalNet) site. In addition, the accuracy of the surface reflectance product generated from the KOMPSAT-3A image, the surface reflectance of Landsat Analysis Ready Data (ARD) of the same site, and near acquisition date were compared with RapidEye-based one. At the same time, a comparative study was carried out with the processing results using QUick Atmospheric Correction (QUAC) and Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes (FLAASH) tool supported by a commercial tool for the same image. Similar to the KOMPSAT-3A-based surface reflectance product, the results obtained from RapidEye Extension showed accuracy of agreement level within 5%, compared with RadCalNet data. They also showed better accuracy in all band images than the results using QUAC or FLAASH tool. As the importance of the Red-Edge band in agriculture, forests, and the environment applications is being emphasized, it is expected that the utilization of the surface reflectance products of RapidEye images produced using this program will also increase.
  • ArticleOctober 31, 2022

    4 5
    Abstract
    This paper presents a novel training data extraction approach using semi-supervised learning (SSL)-based classification without the analyst intervention for time-series land-cover mapping. The SSL- based approach first performs initial classification using initial training data obtained from past images including land-cover characteristics similar to the image to be classified. Reliable training data from the initial classification result are then extracted from SSL-based iterative classification using classification uncertainty information and class labels of neighboring pixels as constraints. The potential of the SSL- based training data extraction approach was evaluated from a classification experiment using unmanned aerial vehicle images in croplands. The use of new training data automatically extracted by the proposed SSL approach could significantly alleviate the misclassification in the initial classification result. In particular, isolated pixels were substantially reduced by considering spatial contextual information from adjacent pixels. Consequently, the classification accuracy of the proposed approach was similar to that of classification using manually extracted training data. These results indicate that the SSL-based iterative classification presented in this study could be effectively applied to automatically extract reliable training data for time-series land-cover mapping.
KSRS
August 2024 Vol. 40, No. 4, pp. 319-418

Most Keyword ?

What is Most Keyword?

  • It is the most frequently used keyword in articles in this journal for the past two years.

Most View

Editorial Office

Korean Journal of Remote Sensing