• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.2
• /
• pp.10-23
• /
• 2009

The fundamental matrix and key frame selection are one of the most important techniques to recover full 3D reconstruction of objects from turntable sequences. This paper proposes a new algorithm that estimates a robust fundamental matrix for camera calibration from uncalibrated images taken under turn-table motion. Single axis turntable motion can be described in terms of its fixed entities. This provides new algorithms for computing the fundamental matrix. From the projective properties of the conics and fundamental matrix the Euclidean 3D coordinates of a point are obtained from geometric locus of the image points trajectories. Experimental results on real and virtual image sequences demonstrate good object reconstructions.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.4
• /
• pp.370-376
• /
• 2023

In the underwater image, it is not clear to distinguish the shape of the target due to underwater noise and low resolution. In addition, as an input of deep learning, underwater images require pre-processing and segmentation must be preceded. Even after pre-processing, the target is not clear, and the performance of detection and identification by deep learning may not be high. Therefore, it is necessary to distinguish and clarify the target. In this study, the importance of target shadows is confirmed in underwater images, object detection and target area acquisition by shadows, and data containing only the shape of targets and shadows without underwater background are generated. We present the process of converting the shadow image into a 3-mode image in which the target is white, the shadow is black, and the background is gray. Through this, it is possible to provide an image that is clearly pre-processed and easily discriminated as an input of deep learning. In addition, if the image processing code using Open Source Computer Vision (OpenCV)Library was used for processing, the processing speed was also suitable for real-time processing.

• Korean Journal of Optics and Photonics
• /
• v.33 no.1
• /
• pp.1-10
• /
• 2022

In this study we introduce a new method for high-spatial-resolution continuous wave (CW) aerosol lidar that has a high spatial resolution in the near field and a low spatial resolution at long distances. A normal lidar system uses a nanosecond-pulse laser and measures the round-trip TOF between the aerosol and laser to obtain range resolution. In this study, however, we propose a new type of spatially resolving aerosol lidar that uses laser-scattering images. Using a laser-light-scattering image, we have calculated the distance of each scattering aerosol image for a given pixel, and recovered the short-range aerosol extinction. For this purpose, we have calculated the distance image and the contribution range of the aerosol to the given one-pixel image, and finally we have calculated the extinction coefficients of the aerosol with range-resolved information. In the case of traditional aerosol lidar, we can only obtain the aerosol extinction coefficients above 400 m. Using our suggested method, it was possible to extend the range of the extinction coefficient lower then several tens of meters. Finally, we can remove the unknown short-range region of pulsed aerosol lidar using our method.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.19 no.3
• /
• pp.129-136
• /
• 2020

In this paper, we propose a thermal imagery-based object detection algorithm for low-light level nighttime surveillance system. Many features selected by Haar-like feature selection algorithm and existing Adaboost algorithm are often vulnerable to noise and problems with similar or overlapping feature set for learning samples. It also removes noise from the feature set from the surveillance image of the low-light night environment, and implements it using the lightweight extended Haar feature and adaboost learning algorithm to enable fast and efficient real-time feature selection. Experiments use extended Haar feature points to recognize non-predictive objects with motion in nighttime low-light environments. The Adaboost learning algorithm with video frame 800*600 thermal image as input is implemented with CUDA 9.0 platform for simulation. As a result, the results of object detection confirmed that the success rate was about 90% or more, and the processing speed was about 30% faster than the computational results obtained through histogram equalization operations in general images.

• Korean Journal of Soil Science and Fertilizer
• /
• v.40 no.3
• /
• pp.189-195
• /
• 2007

Using fractal dimensionality theory proposed by Riew and Sposito (1991), we attempted to analyze quantitatively the characteristics of porous distribution for built-in soils in the mini-lysimeter and artificial seed-bed media. The 2" stainless core soil samples were taken from lysimeter soils. Artificial seed-bed media were compacted in the acrylic core filled with raw materials consisted of cocopeat, zeolite and perlite. N (Constant number of partitioned group size smaller media volumes) and r (Self-similarity ratio) parameters consisting of fractal dimension D=log(N)/log(1/r) were obtained by Excel Programme using the Riew and Sposito's fractal model. The pore distribution of tested media was screened in pore size and its occurring frequency. The results reveal that the distribution range of pores is wider in the lysimeter soils than in the seed-bed media, while average size of pores in the media is smaller in lysimeter core soils than in seed-bed media.

• Journal of Broadcast Engineering
• /
• v.25 no.4
• /
• pp.576-586
• /
• 2020

Analyzing baseball games using sensors such as radars or riders is expensive. In this paper, we propose an algorithm to detect pitch shots and hit shots using baseball video and to generate ball trajectories within hit shots using camera movement. After the pitch shot and the hit shot detection using object detection and optical flow, we generate the transformation relationship between frames and ball locations in the frame, and calculates the ball trajectory. The performance of the proposed method is evaluated for three KBO baseball video sequences, and the detection accuracy and detection rate of pitch shot and hit shot were within 89-95 [%], and the average error for shot range was 13.6[m], The direction error was 7.5° and foul classification accuracy was 98.6%.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.4 no.3
• /
• pp.1-12
• /
• 2001

This thesis represents the integrated 3D DEM using both the process of satellite image and the real value of topographic maps. This DEM is draped on satellite image processed to improve representations of the real world. The 3D visualization and 3D animation with satellite imagery data enables to depict more vivid and realistic world. The paper also describes and implements the natural environmental/ecological information system that consists of 7 modules to manage environmental data systematically through an enhanced user interface. We make use of topographic map, satellite imagery data and several thematic maps. Each module has a user interface enabling to assist particular needs of decision-making for ecological/environmental assessments associated with spatial analysis of ecosystem and classification of the environmental status quo and other purposes.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.1325-1328
• /
• 2022

언제 어디서든 한 손으로 미디어 콘텐츠를 소비할 수 있게 해주는 모바일 기기들이 기존 전통적 미디어 콘텐츠 단말기였던 TV나 데스크톱 PC들을 대체하게 되면서 세로형 영상 콘텐츠에 대한 수요가 나날이 높아져 가고 있다. 이와 더불어 모바일 단말기 제조사들은 서로 간의 경쟁에서 앞서기 위해 제품 차별화 전략을 수립하고 모바일 사용자들의 요구 사항을 세세하게 맞추기 위한 결과, 저마다 다른 디스플레이 해상도 규격을 가진 모바일 기기들이 생산되고 있는 상황이다. 이에 미디어 콘텐츠 제작자들은 기존 가로형 영상 콘텐츠와 더불어 새롭게 요구되는 세로형 영상 콘텐츠들을 저마다 다른 해상도 규격에 맞추는데 많은 시간과 비용을 투자하고 있다. 더 나아가 모바일 단말기 해상도 규격과 맞지 않는 영상 콘텐츠를 시청하게 될 경우, 모바일 사용자 입장에서는 디스플레이 전체 영역을 뷰포트로 잡을 수 없어 시청 만족도가 떨어질 수 있다. 이에 본 논문은 한 번의 콘텐츠 제작을 통해서도 추가 비용 없이 다양한 디스플레이 규격을 가진 단말기들에 대해 맞춤형 콘텐츠 서비스 제공을 가능하게 하여 미디어 콘텐츠 소비자들에게 충분한 시청 몰입감을 제공해줄 수 있는 단말 적응적 미디어 화면비 변환 시스템을 제안한다. 단말 적응적 미디어 화면비 변환 시스템은 딥러닝 네트워크 모델과 이미지 관련 라이브러리를 기반으로 하여 설계한 시스템이며, 사용자가 시청하기 원하는 영역을 판단하고, 사용자가 원하는 뷰포트 종횡비에 따라 해당 영역을 잘라내어 사용자가 원하는 세로형 영상 콘텐츠를 제공해준다.

• Journal of the Korean Society of Radiology
• /
• v.15 no.2
• /
• pp.101-108
• /
• 2021

The CT can accurately present the anatomical structure of an organ in the human body, and the resolution of the image is excellent. On Brain CT examination, the radiation sensitivity of the orbit is high and it is subject to many exposure effects. To reduce exposure dose of lens, this study compares change of exposure dose and shielding rate about non-shielding and shielding in a way of using two shielding materials, bismuth and tungsten. In this study, we used bismuth and tungsten filament as shielding materials made by 3D printing to measure the exposure dose according to the materials thickness and each of slices. To compare each shielding rate, 1 mm to 5 mm of two materials was measured with the head phantom fixed and the Magicmax universal dosimeter placed on the eye when the shielding material is not placed, and the shielding material is placed on it. In the 1 mm thick filament, the bismuth filament showed 26.8% and the tungsten filament showed 43.1% shielding rate. Therefore, tungsten presents much greater shielding effect than bismuth.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.11
• /
• pp.1507-1514
• /
• 2016

Full-image guided filter reflects all pixels of image in filtering by using weight propagation and two-pass model, whereas the existing guide filter is processed based on the kernel window. Therefore the computational complexity can be improved while maintaining characteristics of guide filter, such as edge-preserving, smoothing, and so on. In this paper, we propose an efficient VLSI architecture for the full-image guided filter by analyzing the data dependency, the data frequency and the PSNR analysis of the image in order to achieve enough speed for various applications such as stereo vision, real-time systems, etc. In addition, the proposed efficient scheduling enables the realtime process by minimizing the idle period in weight computation. The proposed VLSI architecture shows 214MHz of maximum operating frequency (image size: 384*288, 965 fps) and 76K of gates (internal memory excluded).