• Korean Journal of Remote Sensing
• /
• v.36 no.6_2
• /
• pp.1591-1604
• /
• 2020

This study is a study on classifying land cover by applying high-resolution satellite images to deep learning algorithms and verifying the performance of algorithms for each spatial object. For this, the Fully Convolutional Network-based algorithm was selected, and a dataset was constructed using Kompasat-3 satellite images, land cover maps, and forest maps. By applying the constructed data set to the algorithm, each optimal hyperparameter was calculated. Final classification was performed after hyperparameter optimization, and the overall accuracy of DeeplabV3+ was calculated the highest at 81.7%. However, when looking at the accuracy of each category, SegNet showed the best performance in roads and buildings, and U-Net showed the highest accuracy in hardwood trees and discussion items. In the case of Deeplab V3+, it performed better than the other two models in fields, facility cultivation, and grassland. Through the results, the limitations of applying one algorithm for land cover classification were confirmed, and if an appropriate algorithm for each spatial object is applied in the future, it is expected that high quality land cover classification results can be produced.

• Korean Journal of Remote Sensing
• /
• v.27 no.4
• /
• pp.445-455
• /
• 2011

A probabilistic classification framework is presented that can combine temporal contextual information derived from an existing land-cover map in order to improve the classification accuracy of land-cover classes that can not be discriminated well when using spectral information only. The transition probability is computed by using the existing land-cover map and training data, and considered as a priori probability. By combining the a priori probability with conditional probability computed from spectral information via a Bayesian combination rule, the a posteriori probability is finally computed and then the final land-cover types are determined. The method presented in this paper can be adopted to any probabilistic classification algorithms in a simple way, compared with conventional classification methods that require heavy computational loads to incorporate the temporal contextual information. A case study for crop classification using time-series MODIS data sets is carried out to illustrate the applicability of the presented method. The classification accuracies of the land-cover classes, which showed lower classification accuracies when using only spectral information due to the low resolution MODIS data, were much improved by combining the temporal contextual information. It is expected that the presented probabilistic method would be useful both for updating the existing past land-cover maps, and for improving the classification accuracy.

• Proceedings of the KSRS Conference
• /
• 1999.11a
• /
• pp.48-53
• /
• 1999

The paper presents an automated method for generating fuzzy rules and fuzzy membership functions for pattern classification from training sets of examples and an application to the land-cover classification. Initially, fuzzy subspaces are created from the partitions formed by the minimum and maximum of individual feature values of each class. The initial membership functions are determined according to the generated fuzzy partitions. The fuzzy subspaces are further iteratively partitioned if the user-specified classification performance has not been archived on the training set. Our classifier was trained and tested on patterns consisting of the DN of each band, (XS1, XS2, XS3), extracted from KITSAT-3 multispectral scene. The result represents that our classification method has higher generalization power.

• Journal of Environmental Impact Assessment
• /
• v.26 no.6
• /
• pp.385-403
• /
• 2017

This study showed the national ecosystem classification for the spatial standards of ecosystems-based approaches to the risk assessments and adaptation plan. The characteristics of climate change risk assessment, implement national adaptation plans, and ecosystem/habitat classification status was evaluated. Focusing on the land cover classification widely utilized as spatial data for the assessments of biodiversity and ecosystem services in the UK and other countries in Europe, the applicability of the national land cover classification for climate change risk assessments was reviewed. Considering the ecosystem classification for climate change risk assessment and establishing adaptation measures, it is difficult to apply rough classification method to the land cover system because of lack of information on habitat trend by categorization. The results indicated that forest ecosystems and agro-ecosystem occupied 62.3% and 25.0% of land cover, respectively, of the entire country. Although the area is small compared with the land area, wetland ecosystem (2.9%), marine ecosystem (0.4%), coastal ecosystem (0.6%), and urban ecosystem (6.1%) can be included in the risk assessments. Therefore, it is necessary to subdivide below the medium classification for the forest and agricultural land, as well as Inland wetland, which has a higher proportion of habitat preference of taxa than land area, marine/coastal habitat, and transition areas such as urban and natural ecosystem.

• ETRI Journal
• /
• v.21 no.4
• /
• pp.41-51
• /
• 1999

This research is an attempt to obtain more accurate land cover information from LANDSAT images. Canonical correlation analysis, which has not been widely used in the image classification community, was applied to the classification of a LANDSAT images. It was found that it is easy to select training areas on the classification using canonical correlation analysis in comparison with the maximum likelihood classifier of $ERDAS^{(R)}$ software. In other words, the selected positions of training areas hardly affect the classification results using canonical correlation analysis. when the same training areas are used, the mapping accuracy of the canonical correlation classification results compared with the ground truth data is not lower than that of the maximum likelihood classifier. The kappa analysis for the canonical correlation classifier and the maximum likelihood classifier showed that the two methods are alike in classification accuracy. However, the canonical correlation classifier has better points than the maximum likelihood classifier in classification characteristics. Therefore, the classification using canonical correlation analysis applied in this research is effective for the extraction of land cover information from LANDSAT images and will be able to be put to practical use.

• Proceedings of the KSRS Conference
• /
• v.2
• /
• pp.647-650
• /
• 2006

Compact Airborne Spectrographic Imager (CASI) hyperspectral imagery was acquired over the Little Miami River Watershed (1756 square miles) in Ohio, U.S.A., which is one of the largest hyperspectral image acquisition. For the development of a 4m-resolution land cover dataset, a hierarchical approach was employed using two different classification algorithms: 'Image Object Segmentation' for level-1 and 'Spectral Angle Mapper' for level-2. This classification scheme was developed to overcome the spectral inseparability of urban and rural features and to deal with radiometric distortions due to cross-track illumination. The land cover class members were lentic, lotic, forest, corn, soybean, wheat, dry herbaceous, grass, urban barren, rural barren, urban/built, and unclassified. The final phase of processing was completed after an extensive Quality Assurance and Quality Control (QA/QC) phase. With respect to the eleven land cover class members, the overall accuracy with a total of 902 reference points was 83.9% at 4m resolution. The dataset is available for public research, and applications of this product will represent an improvement over more commonly utilized data of coarser spatial resolution such as National Land Cover Data (NLCD).

• Korean Journal of Remote Sensing
• /
• v.20 no.4
• /
• pp.253-259
• /
• 2004

The purpose of this study is to present a standardized scheme for providing agriculture-related information at various spatial resolutions of satellite images including Landsat + ETM, KOMPSAT-1 EOC, ASTER VNIR, and IKONOS panchromatic and multi-spectral images. The satellite images were interpreted especially for identifying agricultural areas, crop types, agricultural facilities and structures. The results were compared with the land cover/land use classification system suggested by National Geographic Information based on aerial photograph and Ministry of Environment based on satellite remote sensing data. As a result, high-resolution agricultural land cover map from IKONOS imageries was made out. The classification result by IKONOS image will be provided to KOMPSAT-2 project for agricultural application.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.49 no.5
• /
• pp.67-77
• /
• 2007

The images of MODerate resolution Imaging Spectroradiometer (MODIS) that provide wider swath and shorter revisit frequency than Land Satellite (Landsat) and Satellite Pour I' Observation de la Terre (SPOT) has been used fer land cover classification with better spatial resolution than National Oceanic and Atmosphere Administration/Advanced Very High Resolution Radiometer (NOAA/AVHRR)'s images. Due to the advantages of MODIS, several researches have conducted, however the results for the land cover classification using MODIS images have less accuracy of classification in small areas because of low spatial resolution. In this study, uncertainty of paddy fields classification using MODIS images was conducted in the region of Gyeonggi-do and the relation between this uncertainty of estimating paddy fields and topographical factors was also explained. The accuracy of classified paddy fields was compared with the land cover map of Environmental Geographic Information System (EGIS) in 2001 classified using Landsat images. Uncertainty of paddy fields classification was analyzed about the elevation and slope from the 30m resolution Digital Elevation Model (DEM) provided in EGIS. As a result of paddy classification, user's accuracy was about 41.5% and producer's accuracy was 57.6%. About 59% extracted paddy fields represented over 50 uncertainty in one hundred scale and about 18% extracted paddy fields showed 100 uncertainty. It is considered that several land covers mixed in a MODIS pixel influenced on extracted results and most classified paddy fields were distributed through elevation I, II and slope A region.

• Proceedings of the Korean Association of Geographic Inforamtion Studies Conference
• /
• 2004.03a
• /
• pp.261-265
• /
• 2004

The purpose of this study is to present a standardized scheme for providing agriculture-related information at various spatial resolutions of satellite images including Landsat＋ETM, KOMPSAT-1 EOC, ASTER VNIR, and IKONOS panchromatic and multi-spectral images. The satellite images were interpreted especially for identifying agricultural areas, crop types, agricultural facilities and structures. The results were compared with the land cover/land use classification system suggested by Ministry of Construction & Transportation based on NGIS (National Geographic Information System) and Ministry of Environment based on satellite remote sensing data. As a result, high-resolution agricultural land cover map from IKONOS imageries was made out. The results by IKONOS image will be provided to KOMPSAT-2 project for agricultural application.

• Korean Journal of Remote Sensing
• /
• v.23 no.5
• /
• pp.483-491
• /
• 2007

One of the major obstacles to classify and validate Land Cover maps is the high cost of acquiring reference data. In case of inaccessible areas such as North Korea, the high resolution satellite imagery may be used for reference data. The objective of this paper is to investigate the possibility of utilizing QuickBird high resolution imagery of North Korea that can be obtained from Google Earth data via internet for reference data of land cover classification. Monthly MODIS NDVI data of nine months from the summer of 2004 were classified into L=54 cluster using ISODATA algorithm, and these L clusters were assigned to 7 classes - coniferous forest, deciduous forest, mixed forest, paddy field, dry field, water, and built-up areas - by careful use of reference data obtained through visual interpretation of the high resolution imagery. The overall accuracy and Kappa index were 85.98% and 0.82, respectively, which represents about 10% point increase of classification accuracy than our previous study based on GCP point data around North Korea. Thus we can conclude that Google Earth may be used to substitute the traditional reference data collection on the site where the accessibility is severely limited.