• Korean Journal of Remote Sensing
• /
• v.39 no.6_3
• /
• pp.1771-1777
• /
• 2023

With the advancements of deep learning, many semantic segmentation-based methods for land cover classification have been proposed. However, existing deep learning-based models only use image information and cannot guarantee spatiotemporal consistency. In this study, we propose a land cover classification model using geographical coordinates. First, the coordinate features are extracted through the Coordinate Hash Encoder, which is an extension of the Multi-resolution Hash Encoder, an implicit neural representation technique, to the longitude-latitude coordinate system. Next, we propose an architecture that combines the extracted coordinate features with different levels of U-net decoder. Experimental results show that the proposed method improves the mean intersection over union by about 32% and improves the spatiotemporal consistency.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.23 no.3
• /
• pp.252-262
• /
• 2020

In Remote Sensing, a machine learning based SVM model is typically utilized for land cover classification. And study using neural network models is also being carried out continuously. But study using high-resolution imagery of KOMPSAT is insufficient. Therefore, the purpose of this study is to assess the accuracy of land cover classification by neural network models using high-resolution KOMPSAT-3 satellite imagery. After acquiring satellite imagery of coastal areas near Gyeongju City, training data were produced. And land cover was classified with the SVM, ANN and DNN models for the three items of water, vegetation and land. Then, the accuracy of the classification results was quantitatively assessed through error matrix: the result using DNN model showed the best with 92.0% accuracy. It is necessary to supplement the training data through future multi-temporal satellite imagery, and to carry out classifications for various items.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2012.05a
• /
• pp.117-117
• /
• 2012

본 연구에서는 도시화가 진행중인 경안천($561.1km^2$)유역과 안성천($909.1km^2$) 유역을 대상으로 토지이용 변화를 분석하고 그에 따른 유출특성을 파악하는데 있다. 경안천과 안성천 유역의 과거(1975, 1980, 1985, 1990, 1995, 2000)의 토지이용 분석결과 도시지역 면적이 각각 6.9%와 8.3%씩 증가하였다. 본 연구에서는 미래토지이용 예측을 위하여 CLUE-S (Conversion of Land Use Change and its Effects)모델을 이용하였다. 과거 토지이용 변화 특성을 분석하여 토지피복의 변화와 전이 특성값을 결정하였고, 이를 바탕으로 토지면적 시나리오, 변화 제한지역, 회귀식 결과와 토지이용 변화 특성 값을 CLUE-S 모델에 입력하여 미래의 토지이용변화를 모의하였다. 예측된 토지이용 정보를 가지고 2040s (2020-2059)와 2080s (2060-2099)의 기간으로 나누어 수문모의가 가능한 SWAT (Soil and Water Assessment Tool) 모형에 적용 및 그에 따른 유출특성을 파악하였다. SWAT 모형의 적용성 평가를 위하여 안성천과 경안천 유역을 대상으로 1998~2005의 기간 동안 일별 유출, 자료에 대한 보정 및 검증을 실시하였다. 미래의 토지이용 변화에 따른 유역의 유출특성의 영향은 도시의 비율이 큰 유역에서 더 많이 나타날 것으로 기대된다.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.27 no.1
• /
• pp.115-127
• /
• 2024

This research aimed to construct models with various structures based on the Transformer module and to perform land cover classification, thereby examining the applicability of the Transformer module. For the classification of land cover, the Unet model, which has a CNN structure, was selected as the base model, and a total of four deep learning models were constructed by combining both the encoder and decoder parts with the Transformer module. During the training process of the deep learning models, the training was repeated 10 times under the same conditions to evaluate the generalization performance. The evaluation of the classification accuracy of the deep learning models showed that the Model D, which utilized the Transformer module in both the encoder and decoder structures, achieved the highest overall accuracy with an average of approximately 89.4% and a Kappa coefficient average of about 73.2%. In terms of training time, models based on CNN were the most efficient. however, the use of Transformer-based models resulted in an average improvement of 0.5% in classification accuracy based on the Kappa coefficient. It is considered necessary to refine the model by considering various variables such as adjusting hyperparameters and image patch sizes during the integration process with CNN models. A common issue identified in all models during the land cover classification process was the difficulty in detecting small-scale objects. To improve this misclassification phenomenon, it is deemed necessary to explore the use of high-resolution input data and integrate multidimensional data that includes terrain and texture information.

• Journal of Korean Society for Geospatial Information Science
• /
• v.16 no.1
• /
• pp.33-41
• /
• 2008

Land use/land cover is very important to understand the change in the land cover between specific periods. But as there are number of factors which are responsible for the change in the land cover, it is very difficult to identify the specific factors. Therefore in the study we made an attempt to use the land use strategies quantitatively and conducted simulation study. The input data using the CLUE-S model are the satellite data of 1987 and 2001 from Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM+) and we conducted simulations for 23 years from 1987 to 2010. As a result, the accuracy between the land use map derived from original satellite data and simulation for 2001 was 93.69% and in this reason we could expect land use and land cover in the future.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.17 no.1
• /
• pp.1-10
• /
• 1999

Landuse information is base data being applied as common database in the process of executing urban project service and is very available. But, the progress of business on landuse information building is deficient yet. In this study, we'd like to deal with landuse information projet that is the base of urban project, many landuse information building and the method of extraction of efficient second-information. Also, we built system to apply actively landuse information for urban projectors. And, we'd like to present the model on urban landuse classification system that wouldn't be standard yet by studying the method of urban landuse pattern using landuse in-formation being built. Also, we can evaluate if the model of urban landuse pattern comes up to present landuse. We can expect it is a base data by extracting unsuitable area from present landuse efficiently when we construct landuse project.

• Journal of Korean Society for Geospatial Information Science
• /
• v.4 no.1 s.6
• /
• pp.23-29
• /
• 1996

The purpose of this study is to develope an Object-Oriented Graphic database management system to handle geographic data of geographic information systems. As the result of this study, unstructured vector model was developed to handle geographic data and graphic database management was implemented by object-oriented programming. This study was focused on liking function between graphic data and attribute data, and not focused on network analysis function.

• Journal of Cadastre & Land InformatiX
• /
• v.48 no.1
• /
• pp.55-70
• /
• 2018

The field of cadastre and land administration is susceptible to changes in technological development and requires a robust system to respond. In Mongolia, there is a need for a future-oriented and robust cadastral system to cope with these changes. Therefore, this study chose the Land Administration Domain Model (LADM) which is created by the International Organization for Standardization(ISO), and analyzed for its application to the cadastral system of Mongolia. For the application of LADM for Mongolia, we analyzed four packages and each class of LADM based on characteristics of Mongolia's cadastral system and cases of countries applying LADM. As a result of this analysis, we have completed LADM national profiling in Mongolia and proposed the possibility of applying LADM. In the future, based on this research, application of LADM will play an important role in Mongolian cadastral system where the current data model standardization is not established and is useful as a reference model.

• Journal of Cadastre & Land InformatiX
• /
• v.48 no.2
• /
• pp.51-64
• /
• 2018

The paper reviews the adoption of ISO 19152, Land Administration Domain Model (LADM) for the enhancement of the current Mongolian cadastral system. It can be said that the current cadastral system is developed for the pure purpose of land registration. There is a need for a comprehensive data model for not only this reason but also reflecting the current problems in the Mongolian cadastral system. The LADM was published by the International Organization for Standardization later in 2012 as an International Standard for modeling cadastral and land administration information for the purpose of providing a common vocabulary(ontology) and efficient system development. This study examined possibilities of adopting the LADM to the cadastral system for Mongolia focused on Land Manager system. Data model of the Land Manager was examined against the corresponding LADM classes and as a result, gaps between each data model have been drawn. Lastly we proposed the LADM-based new data model for Mongolian cadastral system ensuring that the current problems be reflected.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.4 no.3
• /
• pp.31-40
• /
• 2001

Geography Information System is divided to many details fields such as Urban Information System, Land Information System, Military Information System etc. These detailed fields are connected each other and make National Geography Information System. Now Geography Information System is being used in many fields with Urban Information System. And information of all field is being constructed to network for share each other. Now Land Information System(LIS) is being constructed to two dimensional. But LIS can construct and utilize three dimensional geographic data by connecting Geography Information System and this effect will be greatest. Thus, the study of connecting cadastral map and digital terrain map must be continued. Through the study of connecting digital terrain map, the construction and analysis of three dimensional digital elevation model will be able to construct Land Information System effectively. To this study, the authors constructed integrated geographic data by uniting digital terrain map and cadastral map and constructed three dimensional digital elevation model. By connecting cadastral information database, the authors developed three dimensional Integrated Land Information System.