• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.26 no.2
• /
• pp.165-173
• /
• 2008

This study aims at generating automatically three dimensional geometric models of roads using LIDAR data and a digital map. The main processes in the proposed method are (1) generating a polygon encompassing a road region using a road layer from the digital map, (2) extracting LIDAR points within the road region using the polygon, (3) organizing the points into surface patches and grouping the patches into surface clusters, (4) searching the road surface clusters and generating the surface model from the points linked to the clusters, (5) refining the boundary using a digital map. By applying the proposed method to real data, we successfully generated the linear and surface information of the roads.

• Journal of Korean Society for Geospatial Information Science
• /
• v.23 no.2
• /
• pp.105-111
• /
• 2015

Domestic 2D maps includes only most up-to-date information at the time of production without historical information. Therefore, it is hard to identify the change history of real world objects. In this research, Spatio-temporal model for 2D map were developed and it's compatibility was verified through the pilot project conducted on the Gwanggyo area of Gyeonggi province. Also, the procedure to generate 2D spatio-temporal database using maps made periodically on the same target area was introduced for showing the possibility of realizing nation wide spatio-temporal 2D map using the national base map updated periodically.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.32-36
• /
• 2020

문장의 의미를 잘 임베딩하는 문장 인코더를 만들기 위해 비지도 학습과 지도 학습 기반의 여러 방법이 연구되고 있다. 지도 학습 방식은 충분한 양의 정답을 구축하는데 어려움이 있다는 한계가 있다. 반면 지금까지의 비지도 학습은 단일 형식의 말뭉치에 한정해서 입력된 현재 문장의 다음 문장을 생성 또는 예측하는 형식으로 문제를 정의하였다. 본 논문에서는 위키피디아, 뉴스, 지식 백과 등 문서 형태의 말뭉치에 더해 지식인이나 검색 클릭 로그와 같은 구성이 다양한 이종의 대량 말뭉치를 활용하는 자기 지도 학습 방법을 제안한다. 각 형태의 말뭉치에 적합한 자기 지도 학습 문제를 설계하고 학습한 경우 KorSTS 데이셋의 비지도 모델 성능 평가에서 기준 모델 대비 7점 가량의 성능 향상이 있었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.798-800
• /
• 2024

현대의 머신러닝 환경에서는 특히 모바일 컴퓨팅 및 사물 인터넷(IoT)의 애플리케이션 영역에서 개인 정보를 보호하고 효율적이며 확장 가능한 모델에 대한 관심이 높아지고 있다. 본 연구는 연합 학습(FL)과 자기지도 학습(self-supervised learning)을 결합하여 이질적(heterogeneous)인 분산 자원에서 레이블이 없는 데이터를 활용하면서 사용자의 개인 정보를 보호하는 새로운 프레임워크를 소개한다. 이 프레임워크의 핵심은 SimCLR 과 같은 자기지도 학습 기법으로 학습된 공유 인코더로, 입력 데이터에서 고수준 특성을 추출하도록 설계되었다. 또한 이 구조를 통해 주석(annotation)이 없는 방대한 데이터셋을 활용하여 모델 성능을 향상시키고, 여러 개의 격리된 모델이 필요하지 않아 리소스를 크게 최적화할 수 있는 가능성을 확인했다. 본 연구를 통해 생성된 모델은 중앙 집중 방식(CL)이면서 자기지도학습으로 학습되지 않은 기존 모델과 비교하여 전체 평균 정확도가 14.488% 향상됐다.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.32 no.4_1
• /
• pp.335-341
• /
• 2014

The main goal of this study is on developing a proper brand-new data of national base map and Data Based(DB) model for new information technology environments. To achieve this goal, we generated a brand-new Hybrid spatial information model which is specialized in the spatio-temporal map structure, the framework map for information integration, and the multiple-layered topology structure. The DB structure was designed to reflect the change of objections by adding a new dimension of 'time' in the spartial information, while the infrastructure was able to connect/converge with other information by giving the unique ID and multi-scale fusion map structure. Furthermore, the topology and multi visualization structure, including indoor and basement information, were designed to overcome limitations of expressing in 2 dimension map. The result from the performance test, which was based on the Hybrid spatial information model, confirms the possibility in advanced national base map and conducted DB model through implementing various information and spatiotemporal connections.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2002.04a
• /
• pp.207-217
• /
• 2002

수치정사사진지도를 제작하기 위한 기반 자료로 스캐닝한 수치항공사진영상과 항공사진의 왜곡 보정을 위한 표정, DEM의 추출과 편집 및 편위 수정에 이르기까지 다양한 처리절차가 필요하다. 또한 영상과 벡티 자료의 중첩을 통해 생선한 영상지도의 편집과 지도제작 작업에는 다양한 미적 요소가 필요하다. 본 연구에서는 수치항공사진 영상을 이용하여, 영상에 내포하는 기하학적 변위를 3차원 모델링 결과와 수치표고모델을 이용하여 소거하고, 수치정사사진을 제작한 후에, 1/5,000 수치지도에서 필요한 레이어와 중첩시킨 수치정사사진지도를 제작하여 그 정확도를 분석하였으며, 각 처리 단계별 문제점의 분석과 개선방안을 제시하였다.

• Journal of the Korean Geographical Society
• /
• v.39 no.5 s.104
• /
• pp.804-818
• /
• 2004

As spatial data resources become more abundant, the potential for conflict among them increases. Those conflicts can exist between two or many spatial datasets covering the same area and categories. Therefore, it becomes increasingly important to be able to effectively relate these spatial data sources with others then create new spatial datasets with matching geometry and topology. One extensive spatial dataset is US Census Bureau's TIGER file, which includes census tracts, block groups, and blocks. At present, however, census maps often carry information that conflicts with municipally-maintained detailed spatial information. Therefore, in order to fully utilize census maps and their valuable demographic and economic information, the locational information of the census maps must be reconciled with the more accurate municipally-maintained reference maps and imagery. This paper formulates a conceptual framework and two map models of map conflation to make geometrically and topologically consistent source maps according to the reference maps. The first model is based on the cell model of map in which a map is a cell complex consisting of 0-cells, 1-cells, and 2-cells. The second map model is based on a different set of primitive objects that remain homeomorphic even after map generalization. A new hierarchical based map conflation is also presented to be incorporated with physical, logical, and mathematical boundary and to reduce the complexity and computational load. Map conflation principles with iteration are formulated and census maps are used as a conflation example. They consist of attribute embedding, find meaning node, cartographic 0-cell match, cartographic 1-cell match, and map transformation.

• Journal of Korean Society for Geospatial Information Science
• /
• v.3 no.2 s.6
• /
• pp.115-130
• /
• 1995

The National Geographic Institute(NGI), a national mapping agency, has begun to digitalize national base maps to vitalize nation-wide GIS implementations. However, the NGI's cartographic database design reflects only paper map production and is considered inflexible for various applications. In order to suggest an appropriate data model and database implementation method, approaches of two mapping agencies are analyzed: the United State Geological Survey and Ordnance Survey in the United Kingdom One important finding from the analysis is that each data model is designed to achieve two production purposes in the same time : map and data. By taking advantageous features from the two approaches, an ideal model is proposed. To adapt the ideal model to tile present situation in Korean GIS community, a realistic model is generated, which is an 'SDTS-oriented' data model. Because SDTS will be a Korean data transfer standard, it will be a common basis in developing other data models for different purposes.

• Journal of Cadastre & Land InformatiX
• /
• v.51 no.1
• /
• pp.39-51
• /
• 2021

Recently, there has been much interest in practical use of standardized HD map that can effectively define roads, lanes, junctions, road signs, and road facilities in autonomous driving. Various kinds of de jure or de facto standards such as ISO 22726-1, ISO 14296, HERE HD Live map, NDS open lane model, OpenDRIVE, and NGII HD map are currently being used. However, there are lots of differences in data modeling among these standards, it makes difficult to use them together in autonomous driving. Therefore, we propose a data model conversion method to enable an efficient use of various kinds of HD map standards in autonomous driving in this study. Specifically, we propose a conversion method between the NGII HD map model, which is easily accessible in the country, and the OpenDRIVE model, which is commonly used in the autonomous driving industry. The proposed method consists of simple conversion of NGII HD map layers into OpenDRIVE objects, new OpenDRIVE objects creation corresponding to NGII HD map layers, and linear transformation of NGII HD map layers for OpenDRIVE objects creation. Finally, we converted some test data of NGII HD map into OpenDRIVE objects, and checked the conversion results through Carla simulator. We expect that the proposed method will greatly contribute to improving the use of NGII HD map in autonomous driving.

• Journal of Korean Society for Geospatial Information Science
• /
• v.25 no.1
• /
• pp.55-61
• /
• 2017

The V-world, korean type spatial information open platform, provides various services to easily utilize 2D, 3D map and administrative information of the country. Among them, V-world 3D map service, modeled in individual building unit, require requests for each building model file and the draw calls for drawing models on the screen by the request. This causes a large number of model requests and draw calls to occur that increase the latency occurring during the transmission and conversion process between the central processing unit(CPU) and the graphic processing unit(GPU), which lead to the performance degradation of the 3D map service. In this paper, we propose a performance improvement plan to reduce the performance degradation of 3D map service caused by multiple model requests and draw calls. Therefore, we tried to reduce the number of requests and draw calls for the model file by applying a 3D tile model that combined multiple building models to single tile. In addition, we applied the quadtree algorithm to reduce the time required to load the model file by shortening the retrieval time of the model. This is expected to contribute to improving the performance of 3D map service of V-world.