• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.6
• /
• pp.57-65
• /
• 2018

Conventional RF signal-based indoor localization techniques such as BLE or Wi-Fi based fingerprinting method show considerable localization errors even in small-scale indoor environments due to unstable received signal strength(RSS) of RF signals. Therefore, it is difficult to apply the existing RF-based fingerprinting techniques to large-scale indoor environments such as airports and department stores. In this paper, instead of RF signal we use the geomagnetic sensor signal for indoor localization, whose signal strength is more stable than RF RSS. Although similar geomagnetic field values exist in indoor space, an object movement would experience a unique sequence of the geomagnetic field signals as the movement continues. We use a deep neural network model called the recurrent neural network (RNN), which is effective in recognizing time-varying sequences of sensor data, to track the user's location and movement path. To evaluate the performance of the proposed geomagnetic field based indoor positioning system (IPS), we constructed a magnetic field map for a campus testbed of about $94m{\times}26$ dimension and trained RNN using various potential movement paths and their location data extracted from the magnetic field map. By adjusting various hyperparameters, we could achieve an average localization error of 1.20 meters in the testbed.

• Journal of Korea Multimedia Society
• /
• v.9 no.4
• /
• pp.421-433
• /
• 2006

This paper proposes a method to insert virtual objects into a real video stream based on feature tracking and camera pose estimation from a set of single-camera video frames. To insert or modify 3D shapes to target video frames, the transformation from the 3D objects to the projection of the objects onto the video frames should be revealed. It is shown that, without a camera calibration process, the 3D reconstruction is possible using multiple images from a single camera under the fixed internal camera parameters. The proposed approach is based on the simplification of the camera matrix of intrinsic parameters and the use of projective geometry. The method is particularly useful for augmented reality applications to insert or modify models to a real video stream. The proposed method is based on a linear parameter estimation approach for the auto-calibration step and it enhances the stability and reduces the execution time. Several experimental results are presented on real-world video streams, demonstrating the usefulness of our method for the augmented reality applications.

• Spatial Information Research
• /
• v.9 no.1
• /
• pp.1-13
• /
• 2001

Nowadays, SAR Interferometry is actively being studied as a new technique in topographic mapping using satellite imagery. It extracts height values using phase information derived by two SAR imageries covering same areas. Unlike when using SPOT imagery, it is not affected by atmospheric conditions and time. So to speak, we can say that SAR Interferometry is flexible in imagery acquisitions and can get height data economically over wide area. So, it is expected that SAR Interferometry will be widely using in GIS applications. But, in some area occurring geometric distortion, height data are misjudged or not extracted depending on phase unwrapping algorithms. IN the case of ERS tandem data, the accuracy of height data was worst in mountain area. It is the because of the short incidence angle resulted in layover effect. Of the phase unwrapping algorithms, path-following was better in height accuracy but could not get data in layover area. In this area, we could get height data using Height Sensitivity. In concludion, we could get DEM that maintained the accuracy of path-following method and have overall data across imagery.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.12C
• /
• pp.744-752
• /
• 2011

In intelligent surveillance system, various methods for detecting abnormal behavior were proposed recently. However, most researches are not robust enough to be utilized for actual reality which often has occlusions because of assumption the researches have that individual objects can be tracked. This paper presents a novel method to detect abnormal behavior by analysing major motion of the scene for complex environment in which object tracking cannot work. First, we generate Visual Word and Visual Document from motion information extracted from input video and process them through LDA(Latent Dirichlet Allocation) algorithm which is one of document analysis technique to obtain major motion information(location, magnitude, direction, distribution) of the scene. Using acquired information, we compare similarity between motion appeared in input video and analysed major motion in order to detect motions which does not match to major motions as abnormal behavior.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.5 s.43
• /
• pp.211-221
• /
• 2006

Manufacturing Execution System(MES) immediately enables users to share the information of systems industrywide, efficiently manages synthetic information with data collection through treating the data in a fast way, and helps their decision-making. MES for real-time information processing requires certain conditions such as data modeling of RFID, which has recently attracted attentions, and monitoring of each product unit from manufacture to sales. However, in the middle of processing the unit with a RFID tag, transponders(readers) can't often read the tag due to reader's malfunctions, intentional damages, loss and the circumstantial effects; for that reason, users are unable to confirm the location of the product unit. In this case, users cannot avoid tracing the path of units with uncertain clues. In this paper we suggest that the unique MES based on RFID and Bayesian Network can immediately track the product unit, and show how to evaluate it.

• Journal of Broadcast Engineering
• /
• v.23 no.6
• /
• pp.939-946
• /
• 2018

In this paper, we implement a viewer program that can monitor effectively using omni-directional images. The program consists of four modes: Normal mode, ROI(Region of Interest) mode, Tracking mode, and Auto-rotation mode, and the results for each mode is displayed simultaneously. In the normal mode, the wide angle image is rendered as a spherical image to enable pan, tilt, and zoom. In ROI mode, the area is displayed expanded by selecting an area. And, in Auto-rotation mode, it is possible to track the object by mapping the position of the object with the rotation angle of the spherical image to prevent the object from deviating from the spherical image in Tracking mode. Parallel programming for processing of multiple modes is performed to improve the processing speed. This has the advantage that various angles can be seen compared with surveillance system having a limited angle of view.

• Fire Science and Engineering
• /
• v.33 no.1
• /
• pp.92-98
• /
• 2019

To determine if water-jet nozzle moves and water jetting are effective according to the location of the fire, this study examined the automatic fire-tracking water cannon system and aan indoor hydrant system, such as water jet centered directivity, water jet range maintainability and water jet shape uniformity. First, an examination to find the center of fire accurately from this system design showed that the water jet centered test was accurate. Second, the water jet range test results showed that when water is jetted at the maximum water jet radius, the water jet shows an inaccurate result but within the allowable tolerance range. Finally, the water-jet shape test result confirmed that there are no problems in setting the block from the algorithm design.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.202-202
• /
• 2020

산지 및 도시에서 발생하는 돌발홍수가 대상인 예경보는 홍수 도달시간이 짧고, 수위가 급격하게 상승하는 특성 때문에 1시간 선행시간 확보를 목표로 한다(MOLIT, 2016). 그러나 현재 돌발홍수 예경보 process에 소요되는 시간은 그 이상으로 확인되고 있다. 또한, 돌발홍수 예경보시스템으로부터 출력된 예측 결과를 사람이 직접 확인해야 한다는 단점도 있다. 본 연구에서는 돌발홍수 예경보 선행시간 1시간 확보를 목표로 backward-forward tracking 기법 기반 예측강우 편의보정기법을 제안하고자 한다. 이 기법은 현재 시점보다 이전에 보정계수를 결정함으로써 돌발홍수 예경보 소요시간을 크게 줄여 돌발홍수 대피시간을 확보할 수 있게 한다. 또한, 보정계수의 결정과 적용이 연속적으로 이루어짐에 따라 10분 간격으로 생성되는 MAPLE의 지속적인 편의보정이 가능하다. 예측강우에 대한 보정계수는 현재보다 10분 이전에 결정한다. 즉, 10분 이전 시점에 생성된 10분, 70분 선행 예측강우에 backward tracking을 적용하여 현재 시점의 호우 위치인 target window를 찾는다. 그리고 target window에서 보정계수를 결정한다. 결정된 보정계수는 돌발홍수발령 대상지역인 correction window의 현재 생성된 60분 선행 예측강우에 적용한다. 이 과정에서 과거 시점 10분 선행 예측강우와 현재 시점에 생성된 60분 선행 예측강우와의 forward tracking이 수행된다. Storm tracking 기법으로는 두 예측강우의 호우패턴에 대한 유사성을 정량화한 패턴상관계수를 이용하였다. 대상 호우사상으로는 2016년에 발생한 주요 호우사상을 선정하였다. 본 연구에서 제안하는 기법을 적용하고, 편의보정 결과를 기존 편의보정기법 적용 결과와 비교하였다.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.6
• /
• pp.165-178
• /
• 2011

A gigantic ship is constructed by assembling various types of ship blocks, each block being made by cutting and piecing the steel-plates together. The steel-plate piling process as the initial stage of ship construction sorts and manages the steel-plates according to the ship blocks that the steel-plates are used to make. The steel-plate piling process poses some problems such as process delay due to piling errors, safety vulnerability due to the handling of extra heavy-weight objects, and the uncertainty of work plan due to lack of information management in the pile spaces. We constructed a steel-plate piling process system based on the context-aware computing to resolve such problems. We built simulation system that can simulate the piling process and then established a smart space within the system by using tags, sensors and a real-time location system in order to collect context information. Workers receive an appropriate or intelligent service from the system.

• Journal of KIISE:Databases
• /
• v.30 no.3
• /
• pp.320-335
• /
• 2003

Location-Based Services(LBS) give rise to location-based queries of which results depend on the locations of moving objects. One of important applications of LBS is to examine tracks of continuously moving objects. Moving objects databases need to provide 3-dimensional indexing for efficiently processing range queries on the movement of continuously changing positions. An extension of the 2-dimensional R-tree to include time dimension shows low space utilization and poor search performance, because of high overlap of index nodes and their dead space. To solve these problems, we propose a new R-tree based indexing technique, namely TR-tree. To increase storage utilization, we assign more entries to the past node by using the unbalanced splitting policy. If two nodes are highly overlapped, these nodes are forcibly merged. It is the forced merging policy that reduces the dead space and the overlap of nodes. Since big line segments can also affect the overlap of index nodes to be increased, big line segments should be clipped by the clipping policy when splitting overfull nodes. The TR-tree outperforms the 3DR-tree and TB-tree in all experiments. Particularly, the storage utilization of the TR-tree is higher than the R-tree and R*-tree.