• Journal of Broadcast Engineering
• /
• v.13 no.4
• /
• pp.544-553
• /
• 2008

With the development of video editing technology, there are growing uses of overlay text inserted into video contents to provide viewers with better visual understanding. Since the content of the scene or the editor's intention can be well represented by using inserted text, it is useful for video information retrieval and indexing. Most of the previous approaches are based on low-level features, such as edge, color, and texture information. However, existing methods experience difficulties in handling texts with various contrasts or inserted in a complex background. In this paper, we propose a novel framework to localize the overlay text in a video scene. Based on our observation that there exist transient colors between inserted text and its adjacent background a transition map is generated. Then candidate regions are extracted by using the transition map and overlay text is finally determined based on the density of state in each candidate. The proposed method is robust to color, size, position, style, and contrast of overlay text. It is also language free. Text region update between frames is also exploited to reduce the processing time. Experiments are performed on diverse videos to confirm the efficiency of the proposed method.

• Journal of Korean Society for Geospatial Information Science
• /
• v.12 no.4 s.31
• /
• pp.45-51
• /
• 2004

Control stations managed by national and local governmes are associated with other survey work and constructing geography information and they are important assets in the national level as the positional standard of the country. Since these control points are managed as some type of register and the control points could not be easily updated due to the loss of control stations from construction work or urban development. Therefore, the users could not understand the present situation of the changed control stations. In this background, the aim of this study was to develop control station management system which the managers can use to efficiently maintain control points and to support the usage of the survey control points. For developing this system, we have designed input, update, network, analysis and statistic functions, and have constructed the system using Mapobject as main engine with other languages such as Visual C++ and Visual Basic. The graphic data used in this system are 1/5,000 digital map and digital cadastral map, and the attribute data of each control station are point name, map tile name, longitude and latitude coordinates, TM coordinates, surveying data with the format of year-month-day and control situation photos and so on. In the result of constructing this control station management system, we could achieve integrated management of graphic, attribute and positioning information of each control station.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.11 no.1
• /
• pp.31-44
• /
• 2008

The 1st phase framework data implementation of National Geographic Information Systems (NGIS) used 1:5,000 digital map with 5 years updating period which is lacking in the latest information. This is a significant factor which hinders the use of framework data. This study proposed the efficient technical method of a location based object data management and system implementation for updating framework data. First, we did an object-oriented data modeling and database design using a location based features identifier(UFID: Unique Feature IDentifier). The second, we developed the system with various functions such as a location based UFID creation, input and output, a spatial and attribute data editing, an object based data processing using UML(Unified Modeling Language). Finally, we applied the system to the study area and got high quality data of 99% accuracy and 35% benefit effect of personnel expenses compare to the previous method. We expect that this study can contribute to the maintenance of national framework data as well as the revitalization of various GIS markets by providing user the latest framework data and that we can develop the methods of a feature-change modeling and monitoring using an object based data management.

• Journal of Korean Society for Geospatial Information Science
• /
• v.19 no.2
• /
• pp.3-8
• /
• 2011

Recently, the demand for enterprise for service map providing and portal site services of a 3D virtual city model for public users has been expanding. Also, accuracy of the data, transfer rate and the update for the update for the lapse of time emerge are considered as more impertant factors, by providing 3D information with the web or mobile devices. With the latest technology, we have seen various 3D data through the web. With the VRML progressing actively, because it can provide a virtual display of the world and all aspects of interaction with web. It offers installation of simple plug-in without extra cost on the web. LiDAR system can obtain spatial data easily and accurately, as supprted by numerous researches and applications. However, in general, LiDAR data is obtained in the form of an irregular point cloud. So, in case of using data without converting, high processor is needed for presenting 2D forms from point data composed of 3D data and the data increase. This study expresses urban LiDAR data in 3D, 2D raster data that was applied by compressing algorithm that was used for solving the problems of large storage space and processing. For expressing 3D, algorithm that converts compressed LiDAR data into code Suited to VRML was made. Finally, urban area was expressed in 3D with expressing ground and feature separately.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.27 no.6
• /
• pp.741-747
• /
• 2009

For construction of 3D virtual city models, airborne digital cameras, laser scanners, multi-oblique photograph systems and other devices are currently being used. With such advanced techniques, precise 3D spatial information can be collected and high quality 3D city models can be built in a considerably large area. The 3D spatial information to be built has to provide the latest information that quickly reflects the causes of any change due to urban development. In this study, a UAV photogrammetric method using low cost UAV and digital camera was proposed to acquire and update 3D spatial information effectively on small areas where information continuously change. In the proposed UAV photogrammetric method, the elements of interior orientation were acquired through camera calibration and the vertical and oblique photographs were taken at 9 points and the 3D drawing of ground control points and buildings was performed using 20 images among the pictured images. This study also analyzed the accuracy of the proposed method comparing with ground survey data and digital map in order to examine whether the method can be used in on-demand 3D spatial information update on relatively small areas.

• Journal of KIISE:Databases
• /
• v.29 no.4
• /
• pp.321-333
• /
• 2002

As web-based services are becoming more and more popular, concurrent updates of spatial data should be possible in the web-based environments in order to use the various services. Web-based GIS applications are characterized by large quantity of data providing and these data should be continuously updated according to various user's requirements. Faced with such an enormous data providing system, it is inefficient for a server to do all of the works of updating spatial data requested by clients. Besides, the HTTP protocol used in the web environment is established under the assumption of 'Connectionless'and 'Stateless'. Lots of problems may occur if the scheme of transaction processing based on the LAN environment is directly applied to the web environment. Especially for long transactions of updating spatial data, it is very difficult to control the concurrency among clients and to keep the consistency of the server data. This paper proposes a solution of keeping consistency during updating directly spatial data in the client-side by resolving the Dormancy Region Lock problem caused by the 'Connectionless'and 'Stateless'feature of the HTTP protocol. The RX(Region-eXclusive) lock and the periodically sending of ALIVE_CLIENTi messages can solve this problem. The protocol designed here is verified as effective enough through implementing in the main memory spatial database system, called CyberMap.

• Proceedings of the Korean Association of Geographic Inforamtion Studies Conference
• /
• 2006.05a
• /
• pp.146-151
• /
• 2006

현재 국가지리정보유통망을 통해 유통되고 있는 수치지도(Digital map)는 대부분 항공 측량사진(Air survey Photographs)이나 위성영상(Satellite images)을 통해 취득된 토지 피복의 형상을 기반으로 제작된 지도들이다. 정사사진으로부터 취득된 지형정보를 기호화, 단순화한 수치지도는 실제 사물의 형태와 많은 차이가 있어 지형과 시설물의 변이를 사실대로 반영하지 못한다. 특히 최근 들어 지하시설물이나 건축기술의 발달로 층별 구조가 다양해진 건축물의 형태 등은 정사사진으로는 식별이 불가능하고 건물의 경우 지붕의 형태로 묘사되는 등 표현에 한계가 있다. 도심지역을 대상으로 제작된 1/1,000수치지형도나 일부 지자체나 공공기관 등에서 제작된 $1/500{\sim}1/2,500$ 수치지형도는 상세한 지형정보를 포함하고 있는 높은 정밀도의 수치지도이지만, 도시지역의 변이가 빠르게 일어나고 있기 때문에 자료 신뢰성과 최신성을 유지하는데 한계가 있다. 따라서 본 연구는 현지조사나 측량에 의존하고 있는 정사사진으로부터 취득이 불가능한 지형자료의 취득이나 수년단위인 갱신주기의 한계를 극복하고 실시간으로 수치지도를 수정갱신 할 수 있는 대안으로 일반측량성과도의 활용방안을 제시하고자 한다. 일반측량업에서 생산되는 일반측량성과도는 임시 도근점에 의해 측량을 실시하고 지적경계점과 현황기지점을 도해법에 의해 지적도와 중첩시켜 제작한 측량도로서, 일반측량 성과도에 절대좌표체계를 부여하고 수치지도와 매칭 기법을 제시함으로서 경제적이고 신속한 실시간 수치지도 수정갱신체계를 확립할 수 있고, 자료의 공유함으로써 중복측량을 방지할 수 있다.보 등)를 활용, 구축된다. 교통분석용 네트워크는 구축시점에 따라 현재 및 장래 네트워크로 구분되며 장래 네트워크는 기준년도부터 2031년까지 5년 단위로 계획된 장래도로를 반영하여 구축된다. 교통주제도 및 교통분석용 네트워크는 국가교통DB구축사업을 통해 구축된 자료로서 교통체계효율화법 제9조의4에 따라 공공기관이 교통정책 및 계획수립 등에 활용할 수 있도록 제공하고 있다. 건설교통부의 승인절차를 거쳐 제공하며 활용 후에는 갱신자료 및 활용결과를 통보하는 과정을 거치도록 되어있다. 교통주제도는 국가의 교통정책결정과 관련분야의 기초자료로서 다양하게 활용되고 있으며, 특히 ITS 노드/링크 기본지도로 활용되는 등 교통 분야의 중요한 지리정보로서 구축되고 있다..20{\pm}0.37L$, 72시간에 $1.33{\pm}0.33L$로 유의한 차이를 보였으므로(F=6.153, P=0.004), 술 후 폐환기능 회복에 효과가 있다. 4) 실험군과 대조군의 수술 후 노력성 폐활량은 수술 후 72시간에서 실험군이 $1.90{\pm}0.61L$, 대조군이 $1.51{\pm}0.38L$로 유의한 차이를 보였다(t=2.620, P=0.013). 5) 실험군과 대조군의 수술 후 일초 노력성 호기량은 수술 후 24시간에서 $1.33{\pm}0.56L,\;1.00{\ge}0.28L$로 유의한 차이를 보였고(t=2.530, P=0.017), 술 후 72시간에서 $1.72{\pm}0.65L,\;1.33{\pm}0.3L$로 유의한 차이를 보였다(t=2.540, P=0.016). 6) 대상자의

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.12 no.6
• /
• pp.629-635
• /
• 2019

In Deep Learning models derivative is implemented by error back-propagation which enables the model to learn the error and update parameters. It can find the global (or local) optimal points of parameters even in the complex models taking advantage of a huge improvement in computing power. However, deliberately generated data points can 'fool' models and degrade the performance such as prediction accuracy. Not only these adversarial examples reduce the performance but also these examples are not easily detectable with human's eyes. In this work, we propose the method to detect adversarial datasets with random noise addition. We exploit the fact that when random noise is added, prediction accuracy of non-adversarial dataset remains almost unchanged, but that of adversarial dataset changes. We set attack methods (FGSM, Saliency Map) and noise level (0-19 with max pixel value 255) as independent variables and difference of prediction accuracy when noise was added as dependent variable in a simulation experiment. We have succeeded in extracting the threshold that separates non-adversarial and adversarial dataset. We detected the adversarial dataset using this threshold.

• Journal of Korean Society for Geospatial Information Science
• /
• v.23 no.2
• /
• pp.39-48
• /
• 2015

Lately, Unmanned Aerial Vehicles(UAV), Unmanned Aerial Systems(UAS) or also often known as drones, as a data acquisition platform and as a measurement instrument are becoming attractive for many photogrammetric surveying applications, especially generation of the high density point clouds(HDPC). This paper presents the performance evaluation of a low-cost rotary wing quadrocopter UAS for generation of the HDPC in a test bed environment. Its performance was assessed by comparing the coordinates of UAS based HDPC to the results of Network RTK GNSS surveying with 62 ground check points. The results indicate that the position RMSE of the check points are ${\sigma}_H={\pm}0.102m$ in Horizonatal plane, and ${\sigma}_V={\pm}0.209m$ in vertical, and the maxium deviation of Elevation was 0.570m within block area of ortho-photo mosaic. Therefore the required level of accuracy at NGII for production of ortho-images mosaic at a scale of 1:1000 was reached, UAS based imagery was found to make use of it to update scale 1:1000 map. And also, since this results are less than or equal to the required level in working rule agreement for airborne laser scanning surveying of NGII for Digital Elevation Model generation of grids $1m{\times}1m$ and 1:1000 scale, could be applied with production of topographic map and ortho-image mosaic at a scale of 1:1000~1:2500 over small-scale areas.

• Journal of Institute of Control, Robotics and Systems
• /
• v.19 no.8
• /
• pp.667-675
• /
• 2013

This paper presents vision-based techniques for underwater landmark detection, map-based localization, and SLAM (Simultaneous Localization and Mapping) in structured underwater environments. A variety of underwater tasks require an underwater robot to be able to successfully perform autonomous navigation, but the available sensors for accurate localization are limited. A vision sensor among the available sensors is very useful for performing short range tasks, in spite of harsh underwater conditions including low visibility, noise, and large areas of featureless topography. To overcome these problems and to a utilize vision sensor for underwater localization, we propose a novel vision-based object detection technique to be applied to MCL (Monte Carlo Localization) and EKF (Extended Kalman Filter)-based SLAM algorithms. In the image processing step, a weighted correlation coefficient-based template matching and color-based image segmentation method are proposed to improve the conventional approach. In the localization step, in order to apply the landmark detection results to MCL and EKF-SLAM, dead-reckoning information and landmark detection results are used for prediction and update phases, respectively. The performance of the proposed technique is evaluated by experiments with an underwater robot platform in an indoor water tank and the results are discussed.