• Journal of Broadcast Engineering
• /
• v.25 no.5
• /
• pp.798-807
• /
• 2020

As technologies using digital camera have been developed, 3D images can be constructed from the pictures captured by using multiple cameras. The 3D image data is represented in a form of point cloud which consists of 3D coordinate of the data and the related attributes. Various techniques have been proposed to construct the point cloud data. Among them, Structure-from-Motion (SfM) and Multi-view Stereo (MVS) are examples of the image-based technologies in this field. Based on the conventional research, the point cloud data generated from SfM and MVS may be sparse because the depth information may be incorrect and some data have been removed. In this paper, we propose an efficient algorithm to enhance the point cloud so that the density of the generated point cloud increases. Simulation results show that the proposed algorithm outperforms the conventional algorithms objectively and subjectively.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.11A
• /
• pp.1066-1072
• /
• 2010

In this paper we proposed a multi event handling surveillance system using multiple PTZ cameras. One event is assigned to each PTZ camera to detect unusual situation. If a new object appears in the scene while a camera is tracking the old one, it can not handle two objects simultaneously. In the second case that the object moves out of the scene during the tracking, the camera loses the object. In the proposed method, the nearby camera takes the role to trace the new one or detect the lost one in each case. The nearby camera can get the new object location information from old camera and set the seamless event link for the object. Our simulation result shows the continuous camera-to-camera object tracking performance.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.30 no.6_1
• /
• pp.511-518
• /
• 2012

It is well-known that non-metric digital cameras have to be calibrated for the close-range photogrammetry. But, the self-calibration is still not an easy task because it requires rather a large calibration site of accurate control points, multiple image acquisitions in different positions, and accurate image point measurements that are quite labor-intensive and time-consuming. Based on the premise, this study carried out check point accuracy analysis from self-calibration of different control point designs and photo combinations. The test results showed that the calibration using three photos covering three-dimensional control points produced high accuracy, but control points on a plane could attain the comparable accuracy with four photos including a 90-degree rotated photo. We then compared the target accuracy of on-site self-calibration using flat control points to that of laboratory-self calibration and observed comparable results.

• ETRI Journal
• /
• v.37 no.6
• /
• pp.1143-1153
• /
• 2015

In this paper, we propose an efficient architecture for a real-time image stitching engine for vision SoCs found in motor vehicles. To enlarge the obstacle-detection distance and area for safety, we adopt panoramic images from multiple telegraphic cameras. We propose a stitching method based on a fixed homography that is educed from the initial frame of a video sequence and is used to warp all input images without regeneration. Because the fixed homography is generated only once at the initial state, we can calculate it using SW to reduce HW costs. The proposed warping HW engine is based on a linear transform of the pixel positions of warped images and can reduce the computational complexity by 90% or more as compared to a conventional method. A dual-core SW/HW image stitching engine is applied to stitching input frames in parallel to improve the performance by 70% or more as compared to a single-core engine operation. In addition, a dual-core structure is used to detect a failure in state machines using rock-step logic to satisfy the ISO26262 standard. The dual-core SW/HW image stitching engine is fabricated in SoC with 254,968 gate counts using Global Foundry's 65 nm CMOS process. The single-core engine can make panoramic images from three YCbCr 4:2:0 formatted VGA images at 44 frames per second and frequency of 200 MHz without an LCD display.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.49 no.1
• /
• pp.20-28
• /
• 2012

Gait analysis is an examination which extracts objective information from observing human gait and assesses the function. The equipments used recently for gait analysis are expensive due to multiple cameras and force plates, and require the large space to set up the system. In this paper, we proposed a method to measure human gait motions in 3D from a monocular video. Our approach was based on particle filtering to track human motion without training data and previous information about a gait. We used dynamics to make physics-based motions with the consideration of contacts between feet and base. In a walking sequence, our approach showed the mean angular error of $12.4^{\circ}$ over all joints, which was much smaller than the error of $34.6^{\circ}$ with the conventional particle filter. These results showed that a monocular camera is able to replace the existing complicated system for measuring human gait quantitatively.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.7
• /
• pp.49-58
• /
• 2011

Augmented Reality(AR) assists human's cognitive ability through the information visualization by substantiating information about virtual situation. This technology is studied in a variety of ways including education, design, industry, and so on, by various supply of information devices equipped with cameras and display monitors. Since the most of the AR system depends on limited interaction that responds to the order from user, it can not reflect diverse real world situation. In this study, we suggest a knowledge based augmented reality system, which is composed of context awareness agent that provides recognized context information, along with knowledge based component that provides intelligent capability by utilizing domain knowledges. With this capability, the augmented object can generate dynamic model intelligently by reflecting context information, and can make the interaction possible among the multiple objects. We developed rule based context awareness system along with 3D model generation, and tested interaction among the augmented objects. And we suggest a framework that can provide a convenient way of developing augmented reality system for user.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.46 no.6
• /
• pp.1-8
• /
• 2009

Detecting entry and exit zones in a view covered by multiple cameras is an essential step to determine the topology of the camera setup, which is critical for achieving and sustaining the accuracy and efficiency of multi-camera surveillance system. In this paper, a graph theoretic clustering method is proposed to detect zones using data points which correspond to entry and exit events of objects in the camera view. The minimum spanning tree (MST) is constructed by associating the data points. Then a set of well-formed clusters is sought by removing inconsistent edges of the MST, based on the concepts of the cluster balance and the cluster density defined in the paper. Experimental results suggest that the proposed method is effective, even for sparsely elongated clusters which could be problematic for expectation-maximization (EM). In addition, comparing to the EM-based approaches, the number of data required to obtain stable outcome is relatively small, hence shorter learning period.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.27 no.2
• /
• pp.261-271
• /
• 2009

Due to the recent rise of a need for 3 dimensional geospatial information on urban areas, general interest in aerial multi-view cameras has been on an increase. The conventional geospatial information system depends solely upon vertical images, while the multi-view camera is capable of taking both vertical and oblique images taken from multiple directions, thus making it easier for the user to interpret the object. Through our research we developed a prototype of a multi-view camera system that includes a camera system, GPS/INS, a flight management system, and a control system. We also studied and experimented with the camera viewing angles, the synchronization of image capture, the exposure delay, the data storage that must be considered for the development of the multi-view camera system.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.7
• /
• pp.1740-1749
• /
• 2009

As mobile devices evolve and digital convergence trend is here to stay, mobile phones are built with multiple functions including cameras, MP3s, TVs and game consoles. As a consequence, such multi-functional mobile phones come to spend more power, facilitating development of next-generation ultra-capacity lithium ion battery. In addition, environmental regulations and rising oil prices cause demand for hybrid cars to keep rising. Accordingly, more and more attention is being paid to medium and large batteries and more efforts are being made to realize lower battery prices, higher outputs and stability. This study presented a patent technology related to the lithium ion battery packing that allows reducing processes related, increasing productivity and recycling parts other than the body. The lithium ion battery pack to which protection circuits are elastically attached provides short circuit protection for the circuit and the body and makes electric connection of the circuit and the body easier.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.4
• /
• pp.164-174
• /
• 2015

In a real scenario, radiometric change has frequently occurred in the stereo image acquisition process using multiple cameras with geometric characteristics or moving a single camera because it has different camera parameters and illumination change. Conventional stereo matching algorithms have a difficulty in finding correct corresponding points because it is assumed that corresponding pixels have similar color values. In this paper, we present a new method based on the local descriptor reflecting intensity, gradient and texture information. Furthermore, an adaptive weight for local descriptor based on the entropy is applied to estimate correct corresponding points under radiometric variation. The proposed method is tested on Middlebury datasets with radiometric changes, and compared with state-of-the-art algorithms. Experimental result shows that the proposed scheme outperforms other comparison algorithms around 5% less matching error on average.